Transistors, the building blocks of modern electronics, are typically made of silicon. Because it’s a semiconductor, this material can control the flow of electricity in a circuit. But silicon has fundamental physical limits that restrict how compact and energy-efficient a transistor can be.

MIT researchers have now replaced silicon with a magnetic semiconductor, creating a magnetic transistor that could enable smaller, faster, and more energy-efficient circuits. The material’s magnetism strongly influences its electronic behavior, leading to more efficient control of the flow of electricity. 

The team used a novel magnetic material and an optimization process that reduces the material’s defects, which boosts the transistor’s performance.

The material’s unique magnetic properties also allow for transistors with built-in memory, which would simplify circuit design and unlock new applications for high-performance electronics.

“People have known about magnets for thousands of years, but there are very limited ways to incorporate magnetism into electronics. We have shown a new way to efficiently utilize magnetism that opens up a lot of possibilities for future applications and research,” says Chung-Tao Chou, an MIT graduate student in the departments of Electrical Engineering and Computer Science (EECS) and Physics, and co-lead author of a paper on this advance.

Chou is joined on the paper by co-lead author Eugene Park, a graduate student in the Department of Materials Science and Engineering (DMSE); Julian Klein, a DMSE research scientist; Josep Ingla-Aynes, a postdoc in the MIT Plasma Science and Fusion Center; Jagadeesh S. Moodera, a senior research scientist in the Department of Physics; and senior authors Frances Ross, TDK Professor in DMSE; and Luqiao Liu, an associate professor in EECS, and a member of the Research Laboratory of Electronics; as well as others at the University of Chemistry and Technology in Prague. The paper appears today in Physical Review Letters.

Overcoming the limits

In an electronic device, silicon semiconductor transistors act like tiny light switches that turn a circuit on and off, or amplify weak signals in a communication system. They do this using a small input voltage.

But a fundamental physical limit of silicon semiconductors prevents a transistor from operating below a certain voltage, which hinders its energy efficiency.

To make more efficient electronics, researchers have spent decades working toward magnetic transistors that utilize electron spin to control the flow of electricity. Electron spin is a fundamental property that enables electrons to behave like tiny magnets.

So far, scientists have mostly been limited to using certain magnetic materials. These lack the favorable electronic properties of semiconductors, constraining device performance.

“In this work, we combine magnetism and semiconductor physics to realize useful spintronic devices,” Liu says.

The researchers replace the silicon in the surface layer of a transistor with chromium sulfur bromide, a two-dimensional material that acts as a magnetic semiconductor.

Due to the material’s structure, researchers can switch between two magnetic states very cleanly. This makes it ideal for use in a transistor that smoothly switches between “on” and “off.”

“One of the biggest challenges we faced was finding the right material. We tried many other materials that didn’t work,” Chou says.

They discovered that changing these magnetic states modifies the material’s electronic properties, enabling low-energy operation. And unlike many other 2D materials, chromium sulfur bromide remains stable in air.

To make a transistor, the researchers pattern electrodes onto a silicon substrate, then carefully align and transfer the 2D material on top. They use tape to pick up a tiny piece of material, only a few tens of nanometers thick, and place it onto the substrate.

“A lot of researchers will use solvents or glue to do the transfer, but transistors require a very clean surface. We eliminate all those risks by simplifying this step,” Chou says.

Leveraging magnetism

This lack of contamination enables their device to outperform existing magnetic transistors. Most others can only create a weak magnetic effect, changing the flow of current by a few percent or less. Their new transistor can switch or amplify the electric current by a factor of 10.

They use an external magnetic field to change the magnetic state of the material, switching the transistor using significantly less energy than would usually be required.

The material also allows them to control the magnetic states with electric current. This is important because engineers cannot apply magnetic fields to individual transistors in an electronic device. They need to control each one electrically.

The material’s magnetic properties could also enable transistors with built-in memory, simplifying the design of logic or memory circuits.

A typical memory device has a magnetic cell to store information and a transistor to read it out. Their method can combine both into one magnetic transistor.

“Now, not only are transistors turning on and off, they are also remembering information. And because we can switch the transistor with greater magnitude, the signal is much stronger so we can read out the information faster, and in a much more reliable way,” Liu says.

Building on this demonstration, the researchers plan to further study the use of electrical current to control the device. They are also working to make their method scalable so they can fabricate arrays of transistors.

This research was supported, in part, by the Semiconductor Research Corporation, the U.S. Defense Advanced Research Projects Agency (DARPA), the U.S. National Science Foundation (NSF), the U.S. Department of Energy, the U.S. Army Research Office, and the Czech Ministry of Education, Youth, and Sports. The work was partially carried out at the MIT.nano facilities.

When a friend buys you a cup of coffee, it’s likely that next time, you’ll return the gesture. This type of reciprocal generosity has been well-documented in behavioral economic studies.

However, anthropologists and other social scientists have known for decades that in the context of relationships where one person has more power, status, or influence, reciprocal generosity is usually not the norm. 

Researchers at MIT have now experimentally demonstrated, for the first time, that small changes to the relationship context can dramatically change people’s actions and expectations of reciprocal generosity. 

During interactions between people of different social status, people tend to expect that generosity will flow one way, and it can be either up or down. It may be that a professor always buys coffee for her students, or that a student always offers to help carry groceries for his resident advisor. Once the precedent is established, it is expected to continue.

One interpretation of the findings is that keeping track of whose turn it is to do a favor is the exception in social interactions, not the rule. That is, it is extra work that we do when we want to maintain equal relationships.

“In many intimate relationships, hierarchical relationships, or other kinds of role-based relationships, you don’t put in the work of trying to keep track of turns,” says Rebecca Saxe, the John W. Jarve Professor of Brain and Cognitive Sciences, a member of the McGovern Institute for Brain Research, and associate dean of science at MIT. “Under this interpretation, we just follow precedent because following a precedent is easier. We all know what to expect, and we don’t have to keep track of what happened last time.”

Saxe is the senior author of the study, which appears in the journal Open Mind. MIT graduate student Alicia Chen is the paper’s lead author.

Changing expectations

Most experimental studies of generosity have been done in the context of behavioral economics and game theory. In such experiments, people are usually paired with a stranger and asked to play games that require coordination. Such studies have found that people tend to use turn-taking and reciprocity as their default strategies. These scenarios, however, are stripped from any social context that might exist between people in the real world.

Saxe and Chen wanted to see if they could measure the effects of social context by incorporating relationships into the type of experiments used to evaluate people’s expectations regarding generosity.

“Where generosity becomes hard and complicated is when it starts to occur in the context of existing relationships, because it changes the terms of the relationships,” Saxe says. “What’s expected of you is very different within a relationship than outside of one.”

To study these effects, the researchers designed experiments in which participants read stories about different types of interactions. In some of the scenarios, the subjects of the stories were described as having either symmetric or asymmetric relationships. In others, they were given specific social relationships such as aunt-niece or manager-employee.

Each story described interactions that might be seen in typical daily life, such as buying coffee for a co-worker or preparing a meal for one’s family. Participants were then asked to predict what would happen the next time the interaction occurred.

In all of these scenarios, the researchers found that people expected that generous acts would be reciprocated when they occurred between individuals in symmetric relationships such as friends, cousins, or co-workers of equal rank. However, their expectations changed for asymmetric relationships, where each person has a different social status. In those cases, people expected that any precedent that was set would continue in the future.

One possible explanation for this is that reciprocity is not the norm but an exception that only occurs in the interactions between equals or strangers, the researchers say. Many of our interactions are with people with whom we have asymmetric relationship, and to maintain those relationships, it’s simply easier to follow precedent.

“If there’s no need to keep track of our equal status, then in some ways it’s the default to fall back on following precedents,” Saxe says.

Maintaining relationships

The study showed that in asymmetric relationships, generosity could flow in either direction. Once that direction was established, it was expected to continue. For example, after an older brother bought concert tickets for a much younger brother, the study participants expected that the older brother would also buy the tickets for the next concert. 

“We found that when people know the relationship is asymmetric, they don’t expect reciprocity; they expect the same action to keep on going,” Chen says. “If the lower-rank person acts generously, people expect that to continue, and if the higher-rank person acts generously, people expect that to continue.”

Following precedents is not only easier, but keeping up these actions may help solidify and define existing relationships. For example, anthropologists have long known that gift-giving helps to construct and maintain social relationships. 

“Following a precedent can be a way of actively maintaining relationships and hierarchies, when the asymmetry of the exchange truly reflects the asymmetry of the relationship,” Saxe says. 

The researchers are now working on creating computational models that could be used to analyze different factors that people take into account when they’re considering whether someone might reciprocate a generous act. In addition to the factors examined in this study, others could include how much each person will benefit, what type of relationship they’re in, and culturally specific expectations of how people should act in different situations.

“One really powerful thing about these models is that we can build in existing theories, add things to the models, and then compare how much these extra factors, like considerations related to social relationships, matter in terms of explaining what people are doing,” Chen says. “This allows us to quantitatively compare the different theories to each other.”

The research was funded by the Simons Foundation Autism Research Initiative and the Patrick J. McGovern Foundation.

For remotely operated underwater vehicles, cloudy and turbulent waters are often a no-go. When vehicles settle on the seafloor or dig through a sandbed, they can kick up clouds of sediment that make it tough for onboard cameras to see through. Often, the only thing to do is to wait until the marine dust settles before a vehicle can safely proceed. 

But a new underwater mapping technique developed by engineers at MIT and the Woods Hole Oceanographic Institution (WHOI) may allow vehicles to see through murky, low-visibility waters. 

The method fuses visual images from optical cameras with acoustic data from sonar sensors. The combination enables a vehicle to quickly map the general shape of its surroundings using sonar, even in low-visibility waters. A vehicle can move toward certain shapes in the sonar-mapped environment, coming close enough for optical cameras to visually resolve specific objects in detail. 

The technique is akin to pairing a dolphin’s echolocation with a sea turtle’s close-range vision to see and navigate through murky water, in real-time. 

The researchers tested the method in tank experiments where they could control the water’s degree of visibility. Even in the cloudiest conditions, the system was able to see through the sediment to map the tank’s environment and visualize centimeter-scale details of objects in the tank. 

The team is further improving the technique, which they’ve named Sonar-MASt3R. They envision that the mapping method could safely guide underwater vehicles through murky environments for a range of applications, including scientific exploration, underwater construction and maintenance, and deep-sea recovery. 

“We hope that this work enables us to do more operations in those challenging, low-visibility environments, and helps provide more coverage in areas that are difficult to operate in today,” says Amy Phung, a graduate student in MIT’s Department of Aeronautics and Astronautics, who led the work. 

Phung presented a paper detailing Sonar-MASt3R this week at the IEEE International Conference on Robotics and Automation (ICRA). The paper’s co-author is Richard Camilli, senior scientist of applied ocean physics and engineering at WHOI. 

The best of both

To see underwater, scientists have generally taken an either/or approach, using either optical cameras or sonar sensors to guide the way. Optical cameras can provide detailed visual imagery of a scene, but only in waters that are relatively clear and well-lit. In contrast, sonar sensors perform just as well in clear and murky water; by emitting acoustic waves and measuring the time and angle at which they return, sonar sensors can determine the exact shape, distance, and depth of objects in the environment, though a sonar map lacks any visual detail. 

To get the best of both modes, scientists have looked to combine the two in a new approach known as “opti-acoustic fusion.” In a handful of prior works, research groups have merged sonar and optical data in mapping techniques that are mostly geared toward object recognition and reconstructing workplace environments. Most techniques require time to sync and process the data and therefore do not work in real-time, while only a few can map an environment in 3D. None have been applied to high-resolution mapping underwater in murky, turbid conditions. 

Phung, who is a student in the MIT-WHOI Joint Program, and Camilli, her advisor, aimed to develop an opti-acoustic fusion technique that would generate detailed 3D maps of underwater environments in real time and in low-visibility conditions. The team was motivated, in part, by challenges in safely recovering unexploded underwater mines.

“There can be old explosives in areas that make it unsafe for ships to be in, and the ability to get rid of those safely is best done by robotics,” Camilli says. “But a lot of these explosives are set in surf zone environments where visibility adds to the challenge of doing this safely. That’s one of many applications that our technique can be used for.”

Cloudy, with a chance of mapping

The new method, Sonar-MASt3R, builds on an existing technique, MASt3R, that was developed by researchers in France. MASt3R is an image matching algorithm that is trained to take in visual images of the same scene and quickly estimate the relative depth of each pixel in the scene. In this way, MASt3R can generate a 3D map of the environment in real-time, based on a camera’s 2D images. 

“The downside is that there is no sense of scale,” Phung says. “It will say ‘this pixel is five units closer than this pixel,’ but it can’t say whether that’s 5 meters or 5 feet.”

Luckily, sonar provides absolute measurements of scale. The timing of sonar reflections can be translated directly into a specific depth and distance of objects that the signals bounced off, as well as their shape and contour. 

In their new work, Phung and Camilli used sonar data to correct MASt3R’s scaling and generate precise 3D maps of underwater environments. Even in murky water, the method’s sonar-corrected map would enable a vehicle to know the precise location of objects, and therefore how far to safely move in for a closer inspection, which the vehicle could then do using conventional optical cameras.

The team tested Sonar-MASt3R in experiments with a tank that they filled with water, sediment, and a variety of objects such as a small boulder, a coffee mug, and a packing crate. Inside the tank, they also set up a robotic arm, onto which they mounted an underwater camera, and a sonar sensor. 

For each experimental run, they first carried out a sweep trajectory, in which the robotic arm slowly swept from one side of the tank to the other to capture sonar and visual data. With this first sweep, Sonar-MASt3R quickly creates a coarse sonar-based map of the shapes and contours of the tank and its objects. The coarse map is then used to record close-up camera images of the objects, which are used to improve the map resolution. A “keyframe” approach quickly compares each new image frame to the last keyframe. If a frame provides new information not contained in the last keyframe, the image is added as a new keyframe to the map. If it is similar, it is immediately discarded. In this way, the approach can quickly fill in the map with relevant visual detail, in real-time. 

The researchers tested their new approach underwater, testing eight different levels of turbidity, which they created by stirring up the tank’s sediment. Compared with other opti-acoustic fusion approaches, Sonar-MASt3R generated more accurate 3D maps and resolved smaller, centimeter-scale details, and in cloudier conditions. In the cloudiest condition, which the robotic arm’s cameras could not see through, its sonar sensors were able to generate a rough map of the tank’s hidden objects. This initial map enabled the arm to move safely through the murk and closer to specific objects, which its underwater camera could then visualize in more detail. 

“An analogy would be if you were to go into a china shop in the dark, and try to pick your way around to find a specific coffee mug without knocking things over,” Camilli offers. “This would allow you to do that.”

The team plans to test the approach in natural underwater conditions, where they suspect that the mapping task should be more straightforward. 

“In a tank, it’s like an echo chamber,” Camilli says. “It’s like trying to do this in a funhouse mirror setting where you get all these distortions and reverberations and ghost images that really complicates the processing. If you put it in the real world, it should be easier.”

Then, they say, Sonar-MASt3R could help scientists safely explore in cloudy, turbid, and murky underwater regions.

“The real value in this effort is so we can use this technology in mission scenarios that are untractable right now,” Phung says. “And there are plenty of untractable missions because we don’t have the observational or perception capabilities.”

This research was supported, in part, by NASA, and the National Science Foundation.

Myriam Heiman, the John and Dorothy Wilson Professor of Neuroscience at MIT, will become the director of MIT’s Picower Institute for Learning and Memory, effective July 1. She succeeds Picower Professor Li-Huei Tsai, who is stepping down after leading the institute for 16 years.

Heiman, a molecular neurobiologist and geneticist, studies the neurodegenerative diseases of the brain’s basal ganglia, including Huntington’s disease and Parkinson’s disease. Using cutting-edge techniques, including single-cell genomics and a powerful transcriptomic technique she helped invent, called translating ribosome affinity purification, she aims to understand the molecular changes that eventually lead to cell death in these diseases. 

“Myriam is an extraordinary scientist, a proven leader within MIT, and a deeply caring and generous mentor. Her research to determine why specific brain cell types are particularly vulnerable to diseases such as Huntington’s has produced studies that are both deep in their insight and sweeping in their scope,” says Nergis Mavalvala, dean of the MIT School of Science and the Curtis and Kathleen Marble Professor of Astrophysics. “I firmly believe that Myriam will be an excellent leader during the Picower Institute’s next chapter.”

“I am honored to take on this role to support the institute’s exceptional scientists and trainees as they pursue discoveries that deepen our understanding of the brain and improve human health,” says Heiman, a professor in MIT’s Department of Brain and Cognitive Sciences (BCS). 

The Picower Institute is a community of 16 neuroscience labs dedicated to understanding the mechanisms that drive learning and memory and related functions such as cognition, emotion, perception, behavior, and consciousness. Institute neuroscientists explore the brain and nervous system at multiple scales, from genes and molecules to cells and synapses to circuits and systems, producing novel insights into how disruptions in these mechanisms can lead to developmental, psychiatric, or neurodegenerative disease. 

Picower Professor Susumu Tonegawa founded the institute as a center in 1994 before a transformative gift from Barbara and Jeffry Picower enabled it to become an institute in 2002. Li-Huei Tsai has served as director since 2009, but announced in March that she would step down after more than 16 years to focus on her research.

Heiman joined the Picower Institute, BCS, and the Broad Institute of Harvard and MIT in 2011, after completing her postdoctoral training at The Rockefeller University. She holds a PhD from Johns Hopkins University and a BA from Princeton University. 

“Ever since joining the institute, Heiman’s research has been guided by the principle that fundamental understanding can lead to breakthroughs in addressing disease,” Tsai says. “Myriam has made it her mission to address these kinds of urgent questions in neuroscience.”

Heiman employs sophisticated DNA and RNA analysis technologies to gain detailed insights into how brain cell states change amid disease, revealing molecular pathways that contribute to the particular vulnerability of different cell types. In 2020, Heiman published the results of an innovative in vivo screening of every mouse gene’s impact on the survival of neurons in the brain, identifying hundreds necessary for sustaining neurons and highlighting a specific gene that promoted their resilience in the context of Huntington’s disease. 

Other studies, both in mice and in postmortem human brain samples, have revealed errant immune responses in neurons and in the brain’s blood vessels that contribute to the disease’s progression. The latter finding arose in a 2022 paper, published with MIT Computer Science and Artificial Intelligence Laboratory colleague Manolis Kellis, that also provided the field one of the first cellular atlases of the brain’s vasculature.  

Her research has also produced insights into other neurodegenerative and psychiatric disorders, including ALS and frontotemporal dementia. In 2024, together with Kellis, Heiman published a paper in Cell showing the diseases have remarkable overlaps at the cellular and molecular levels, revealing potential targets that could yield therapies applicable to both disorders. Heiman’s latest research is also producing new insights into substance use disorders and schizophrenia.

Her research program has garnered many awards. In 2021, Heiman became co-recipient of a National Institutes of Health Transformative Research Award, which “promotes cross-cutting, interdisciplinary approaches that could potentially create or challenge existing paradigms” as part of the NIH’s High-Risk, High-Reward Research program. The next year she also received a prestigious NIH R35 grant to find early triggers of disease progression.

Heiman is also a dedicated teacher and mentor. In 2017, she earned the Department of BCS award for excellence in graduate mentoring; and in 2020, she received the department’s award for excellence in undergraduate teaching. In 2024, she was named one of 23 faculty across MIT who are “committed to caring” — an award given out by MIT’s Office of Graduate Education to faculty members who have served as exceptional mentors to graduate students.

Beyond MIT, Heiman serves on editorial boards and the scientific advisory board of the nonprofit Huntington’s Disease Foundation, an organization that supports research aimed at finding treatments and a cure for Huntington’s and related disorders.. 

Heiman says she is looking forward to her new role in service to MIT by leading the Picower Institute.

“I approach this role with humility and enormous enthusiasm,” Heiman says. “The Picower Institute has an extraordinary legacy, and I’m eager to do everything I can to help support the next generation of transformative research.”

A new kernel, or core program within an operating system, gives researchers a cleaner view of what’s happening inside a processor. Called Fractal and developed at MIT, the kernel has already surfaced previously unknown behavior in Apple’s M1.

When security researchers want to understand what a modern processor is really doing with the kind of detail that determines whether attacks like Spectre and Meltdown are possible, they usually run their experiments on top of an operating system that was never built for the job. They open up macOS or Linux, patch the kernel by hand, and hope the modifications hold. The approach is unstable, hard to reproduce, and on Apple’s platforms, slated for deprecation.

A team at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) decided to build something different. Fractal, an operating system kernel written from the ground up, treats the hardware itself as the object of study. Its first major use, a deep look at branch predictors — a CPU’s way of guessing what code to run next, before it knows for certain, so it doesn’t have to waste time waiting to find out — inside Apple’s M1 processor, has already turned up findings that prior work missed, including the first evidence that a class of speculative attack known as “Phantom” affects Apple Silicon.

“We’re using hardware in ways it wasn’t designed for,” says Joseph Ravichandran, the MIT PhD student in electrical engineering and computer science (EECS) who led the project. “It’s not even obvious that this is a possible thing you could do with the hardware. But we found a way to pull all these different primitives off. It’s like a microscope. If you’ve got a hand magnifying glass, you can see a little bit. But if you had an electron microscope, now we’re really talking. That’s what Fractal is. The electron microscope of operating systems.”

A clean room for chip research

The core problem Fractal solves is one that researchers have worked around for years. Modern processors keep state in many internal structures: branch predictors, caches, translation lookaside buffers, and more. To study how those structures behave across the boundary between user code and kernel code, two domains the chip is supposed to keep isolated, researchers need to run nearly identical experiments on each side of that boundary. On a general-purpose operating system, that is very difficult. The system itself manages privilege levels, address spaces, and scheduling, and it injects its own activity into every measurement.

Fractal inverts the model. It boots directly on bare metal, with no other software running, and exposes primitives that let a single experiment switch privilege levels at runtime while executing the same instructions in the same address space. The team calls the underlying technique multi-privilege concurrency, and it relies on a new construct they introduced: the outer kernel thread, which sits inside a user process’s memory but executes with kernel privileges.

The result is an experimental setup with almost no background noise. Where measurements taken under macOS or Linux are blurred by interrupts, scheduler activity, and address-space management, Fractal produces flat baselines and clean signals.

What Fractal found on the M1

Apple’s M1 implements an ARM specification called CSV2, which is supposed to prevent code running in one privilege level from steering speculation in another. Using Fractal, the MIT team confirmed that the protection works for the execute stage of indirect branch prediction: a user-mode program cannot make the kernel speculatively execute a chosen target through the indirect branch predictor.

But the team also found something the chip’s designers may not have intended. The CPU still fetches the target into the instruction cache before the protection kicks in. That fetch is observable through a side channel, which means user code can still influence what the kernel pulls into its caches across the privilege boundary. The same pattern appeared between processes assigned different address space identifiers.

The team also produced the first evidence that Apple Silicon exhibits Phantom speculation, a class of misprediction previously demonstrated only on AMD and Intel processors. In Phantom, ordinary instructions, including a no-op, can be misinterpreted by the CPU as branches, triggering speculative behavior the program never asked for. On the M1, Fractal showed that Phantom fetches succeed across both privilege levels and address spaces, though the execute phase remains blocked.

A separate Fractal experiment overturned a finding from earlier work on the M1’s conditional branch predictor, which had reported that cross-privilege training worked on Apple’s performance cores, but not its efficiency cores. The Fractal team showed that the conditional branch predictor has no privilege isolation at all, on either core type, and that the earlier result was likely an artifact of macOS quietly migrating threads between cores during system calls.

“For us, it is a true independent variable,” Ravichandran says. “You change the privilege level, nothing else changes. The only thing that could explain whether the attack succeeds or not is the privilege level.”

A tool, not a one-off

Fractal supports x86_64, ARM64, and RISC-V, and consists of more than 31,000 lines of code. The team designed it as infrastructure rather than as a single experiment, with familiar POSIX system calls, a C library, and ports of standard tools like vim, GCC, and the dash shell, so that researchers can move existing experiment code over with minimal friction.

The MIT team disclosed its M1 findings to Apple’s product security team. In an unusual reversal, Apple’s engineers also examined Fractal. 

The longer-term ambition is bigger than any single result. Ravichandran wants Fractal to become to microarchitecture research what tools like QEMU and FFmpeg are to their fields: shared infrastructure that the whole community builds on. 

“My hope is that our results as a community get significantly more reliable, significantly more accurate,” says Ravichadran. “With this reduced noise, this clarity, and this guarantee that you’re running on the right core, on the right system.”

“Fractal is a strong architecture contribution because it turns an often ad hoc microarchitectural reverse-engineering workflow into reusable research infrastructure,” says University of Southern California assistant professor Mengyuan Li, who wasn’t involved in the paper. “By reducing software noise and giving researchers tighter control across privilege boundaries, it makes difficult hardware experiments much easier to interpret.”

Ravichandran worked with Mengjia Yan, an MIT associate professor of EECS and CSAIL principal investigator, on the paper. Their work was supported, in part, by the National Science Foundation, the U.S. Air Force Office of Scientific Research, and ACE, which is part of a program sponsored by the U.S. Defense Advanced Research Projects Agency. They presented their work at the IEEE Symposium on Security and Privacy in San Francisco, California.

MIT professor of physics Pablo Jarillo-Herrero is among 10 researchers worldwide to receive this year’s prestigious Kavli Prize. 

Jarillo-Herrero is co-recipient of the 2026 Kavli Prize in Nanoscience “for foundational work that established the field of twistronics.” His co-recipients are professors Eva Y. Andrei at Rutgers University and Allan MacDonald from the University of Texas at Austin.

These three physicists are being honored for the theoretical foundation and experimental validation of a new field of “twistronics,” where superconductivity, magnetism, and other properties can be obtained by rotating two-dimensional materials such as graphene to a “magic angle.”

A partnership among the Norwegian Academy of Science and Letters, the Norwegian Ministry of Education and Research, and the Kavli Foundation, the Kavli Prizes are awarded every two years to “honor scientists for breakthroughs in astrophysics, nanoscience and neuroscience that transform our understanding of the big, the small and the complex.” The laureates in each field will share $1 million.

“Pablo’s groundbreaking research has once again been given well-deserved recognition,” says Nergis Mavalvala, dean of the MIT School of Science and the Curtis and Kathleen Marble Professor of Astrophysics. “Pablo and his co-recipients have pioneered twistronics, very fundamental scientific research that has opened up a new field with myriad possibilities for novel quantum materials.”

In 2009, using scanning tunneling microscopy and spectroscopy on graphene, most commonly found as a single layer of carbon atoms arranged in hexagons resembling a honeycomb structure, Andrei and her research group demonstrated that small variations in twist angle profoundly modified the electronic structure. This demonstration — that geometric control, rather than chemical composition, could modify a material’s electronic structure — represented a fundamental advance in materials design and arguably launched the field now known as “twistronics.”

In 2011, MacDonald quantitatively explained the emergence of this electronic structure by geometries at discrete magic angles. This framework has since become the theoretical foundation of what are known as moiré materials, and has guided subsequent experimental and theoretical developments across a wide range of twisted and layered systems. 

In 2018, Jarillo-Herrero’s group observed correlated insulating phases and superconductivity in magic-angle twisted bilayer graphene devices. The resulting platform, “combining atomic-scale structural simplicity with electronic tunability, has enabled systematic investigations has had broad and lasting impact across nanoscience and quantum material research,” according to the Kavli Prize citation.

“It was a big surprise, because the technique we used, though conceptually straightforward, was hard to pull off in the lab,” said Jarillo-Herrero recently. He is also the Cecil and Ida Green Professor of Physics at MIT and a member of the Research Laboratory of Electronics. 

“I’m humbled and incredibly honored to be sharing this award with [Andrei and MacDonald],” Jarillo-Herrero noted in an essay describing his journey to the Kavli Prize. “I want to also emphasize that this award honors fundamental physics research in nanoscience. It is incredibly important for society to continue to support fundamental research: Although it often doesn’t have a direct near-term application, in the long run it happens to be the most transformative and impactful in society.”

“Pablo’s research has helped spark a revolution in condensed matter physics and nanoscience, inspiring physicists worldwide to explore superconductivity and other emergent phenomena in engineered quantum materials. This work could potentially lead to the creation of superconductors at room temperature, which would would have an enormous technological impact,” says Deepto Chakrabarty, physics department head and William A. M. Burden Professor in Astrophysics.

Jarillo-Herrero's win brings the number of all-time MIT faculty recipients of the Kavli Prize to nine. Prior winners include Nancy Kanwisher in neuroscience (2024), Bob Langer in nanoscience (2024), Sara Seager in astrophysics (2024), Rainer Weiss in astrophysics (2016), Alan Guth in astrophysics (2014), Mildred Dresselhaus in nanoscience (2012), Ann Graybiel in neuroscience (2012), and Jane Luu in astrophysics (2012).

Interpreting medical ultrasound images is a difficult task, requiring a technician to look at 2D images and mentally arrange them into a 3D representation of what the tissue looks like. 

To make that job easier, MIT researchers developed a new approach to ultrasound imaging that allows the user to visualize a 3D augmented-reality image of the object being scanned. Using a virtual-reality headset, they can see a precise 3D digital representation of what the object actually looks like, making it easier to identify and analyze.

This technique could help speed up the training process for ultrasound technicians and other health care providers who use ultrasound. It could also be deployed for use in hospitals, for tasks such as using ultrasound to place a needle in the right location for a biopsy.

“For training, this could make ultrasound more intuitive and more understandable. On the clinical side, it could be less time-consuming, more accurate, and also give health care providers more peace of mind. They wouldn’t have to wonder if they missed anything,” says Canan Dagdeviren, an associate professor of media arts and sciences at MIT and the senior author of the study.

MIT graduate students Jason Hou and Shrihari Viswanath are the lead authors of the paper, which appears today in Nature Communications Engineering. Other authors of the paper include Bowen Wu ’24 and two MIT Summer Research Program students, Cinay Dilibal, a senior at Dartmouth College, and Tanisha Shende, a senior at Oberlin College.

3D representations

Ultrasound imaging works by bouncing high-frequency sound waves off tissues in the body, which are then reflected back to an ultrasound transducer. The transducer converts these sound waves to electrical signals, which are used to create a 2D image of the tissue. Ultrasound technicians are trained to convert these images into a 3D mental representation of the tissue.

“It's a difficult skill to master, and there are long learning curves,” says Hou. “The hardest thing is this mental tomography bottleneck where you’re trained to reconstruct the 2D slices in your 3D mental space. That is a cognitive burden that can lead to inaccuracies in scanning.”

To reduce that cognitive load, the MIT team thought it could be helpful to combine two technologies: 3D ultrasound imaging and augmented reality (AR). 

Three-dimensional ultrasound imaging is occasionally used in fields such as fetal imaging and echocardiography, which is used to image the heart, but most 3D ultrasound imaging systems are expensive and not widely available. For this study, the MIT team used a real-time 3D system they developed recently for use in breast-cancer detection.

Their new system includes an ultrasound probe, slightly smaller than a deck of cards, that transmits information using a chirped data acquisition system (cDAQ). The probe contains an ultrasound array arranged in the shape of an empty square, a configuration that allows the array to take 3D images of the tissue below.

Because this system has fewer ultrasound elements than a typical 3D ultrasound system, it requires less power and is less expensive to build.

The data collected by the ultrasound probe can then be compressed and streamed into a 3D computer graphics engine called Unreal Engine, which converts the voxel data from the ultrasound image into a direct 3D representation of the object, with no loss of information. Wearing an AR/VR headset, the user can see this 3D rendering representing the internal structure, superimposed over the object’s actual location — like X-ray vision. By tilting their head or approaching from a different direction, the user can see different views of the object, making it easier to identify.

Easier to use

The researchers tested their new technology, which they call AR-VIU (augmented real-time volumetric imaging in ultrasound), with a group of 18 participants. Nine of the subjects were experts in ultrasound technology (including sonographers and physicians), and nine had never used ultrasound before.

Each user performed identification tasks using four different ultrasound technologies. In one condition, they viewed 2D images on a regular screen, which is the way that most ultrasounds are now performed. They also viewed 3D images on a regular screen, as well as two augmented reality conditions: one 2D and one 3D (AR-VIU).

In one round of experiments, users were asked to identify an object embedded in gelatin — such as a spring, a ball, or a screw — inside an opaque container that was scanned with ultrasound. In a second set, they were asked to use a pen to mark the location of “tissue phantom” — a gel-like material engineered to mimic human tissue. This simulates the task of locating the right spot for a needle during a biopsy.

The researchers found that the AR-VIU system significantly improved all users’ ability to identify and locate objects. The effect was especially strong for novices, who performed nearly as well as experts when using AR-VIU. When using the traditional 2D imaging system, experts performed much better than novices.

“Overlaying images with the anatomy and providing 3D visual context makes ultrasound significantly easier for novices to understand,” Viswanath says.

In interviews after the experiments, most of the novices reported that they preferred the AR-VIU approach, with many saying that it made the tasks easier.

“The 3D system imposes less brain drain, it’s more intuitive, and it’s easier to understand what is happening in the targeted region,” Dagdeviren says.

Many of the experts said they preferred the traditional 2D imaging because that is what they were accustomed to and had been trained to use. However, those experts also said they could see the benefits of the AR-VIU system in some situations, such as placing a needle for a biopsy or visualizing the movement of the heart wall during echocardiography.

The researchers are now working on further improving the resolution of the imaging and doing additional tests to demonstrate the accuracy of the AR-VIU technology.

The research was funded by the MIT Media Lab Consortium, the National Science Foundation, an MIT HEALS graduate fellowship, and an MIT-Tata graduate fellowship.

The rise of artificial intelligence is riding on the back of an enormous data center expansion. Data centers are projected to account for anywhere from 9 to 17 percent of total electricity usage in the U.S. by the end of the decade. Today, around a third of data center electricity is devoted to cooling the chips that run AI models.

That’s the process Ferveret is working to make more efficient. The startup, founded by Reza Azizian, a former MIT postdoc in nuclear engineering, and Matteo Bucci, MIT’s Esther and Harold E. Edgerton Associate Professor in the Department of Nuclear Science and Engineering, is adapting an approach from nuclear reactors to cool chips using no water and significantly less electricity.

The company’s cooling system submerges computer servers in a specialized liquid that absorbs heat much more efficiently than air from a fan. What makes the solution different from other liquid cooling systems are the bubbles: Ferveret’s Adaptive Phase Cooling (APC) solution produces much smaller bubbles at the surface of the server, which detach more frequently, accelerating the heat transfer process.

Ferveret is already testing its solutions with companies including CleanSpark, the data center developer and operator, as well as FuriosaAI, an AI accelerator company, and Switch, one of the largest data center operators in the U.S.

In a recent study in collaboration with the Samueli Computer Science Department at the University of California at Los Angeles, Ferveret found its APC solution led to a 15 percent improvement in computational power efficiency compared to state-of-the-art liquid cooling solutions. By combining those savings with Ferveret’s power control system to optimize operating conditions, the company says it allows data centers to get 35 percent more tokens — small pieces of text or data — from their AI models with the same amount of power.

“Our goal is to make data centers as sustainable as possible and help them use every single watt of power to generate tokens, which are the most useful outputs,” Azizian says. “Our system enables the operation of more powerful chips, it helps data centers waste a lot less energy, and it accomplishes all that with zero water consumption.”

From nuclear reactors to AI

Azizian was a postdoc at MIT in 2013 when he met Bucci, who was then a research scientist. They worked on heat transfer in nuclear reactors before Azizian went into industry, where he shifted his focus to cooling chips. Azizian first worked on Microsoft’s HoloLens augmented reality headset and then joined Nvidia, which produces the graphical processing units companies use to train and run the latest AI models. Meanwhile, Bucci continued conducting research at MIT, becoming an assistant professor in 2016.

Azizian walked into his first data center in 2017, where he was struck by the massive, noisy fans that filled the building as they cooled.

“I thought, ‘Holy crap, this is not how you cool facilities,’” Azizian recalls, noting air cooling can still take up 40 percent of the power going into a data center. “It was not an efficient way of doing things, but since it wasn’t hurting the performance, no one cared that the cooling technology was 50 years old.”

Azizian began talking with Bucci about applying their knowledge around optimizing heat transfer in nuclear reactors to data centers. Scientists have spent decades finding better ways to move heat in nuclear reactors.

“Heat transfer determines how much energy you can extract from the reactor core, which translates directly to revenue,” Azizian explains.

The founders started Ferveret in 2021. A lot has changed since Azizian walked into his first data center. Chip companies have packed more and more components onto their chips as the explosion in artificial intelligence has put a premium on squeezing as much computing capacity as possible out of limited power supplies.

That has driven data center operators to use liquid to cool chips — often through a technique known as immersion cooling that submerges chips in liquid. The most effective form of immersion cooling brings the liquid to a boil.

“Liquid is a better heat transfer medium than air. That’s why when you stick your hand into room temperature water it still feels cold,” Bucci explains. “When liquid is boiling, it becomes even better at removing heat because the phase change requires a lot of energy, which is the energy you remove from the chip. That lets you transfer large quantities of heat with minimal temperature differences between the chips and the liquid.”

Unfortunately, boiling liquid adds complexity to the system because it forces operators to capture and reliquefy the bubbles while controlling for pressure, temperature, and fluid inventory.

Ferveret’s system is adapted from a process in nuclear reactors called subcooled boiling. It uses a liquid with a low boiling point and none of the toxic PFAS “forever chemicals” that other approaches rely on. At the surface of the chip, Ferveret’s liquid produces smaller bubbles than other immersion cooling approaches. Those bubbles detach more frequently and quickly recondense in the surrounding liquid, accelerating the bubble-rewetting cycle at the surface of the chip to hasten heat transfer.

Ferveret delivers its APC system in small boxes, each of which houses one server. The founders say their modular systems make it easier to deploy the system and simplify maintenance.

“The physics enable us to get to form factors that weren’t possible in the past,” Azizian says. “Most immersion cooling solutions are large tanks that people submerge the servers in. We have a smaller, modular rack-mounted solution that makes it adaptable to the current infrastructure, so it’s easier for people to deploy our technology.”

Ferveret also offers control software that adjusts the power going to each server in real-time to further improve efficiency.

“We deliver full-stack systems that include the cooling box, the rack, the cooling distribution units, and sensors that measure the temperature and pressure,” Bucci says. “Our software monitors those sensors and optimizes the operating condition inside each box to ensure that energy consumption is minimized in the system.”

AI with fewer resources

In addition to helping data centers to run more efficiently, Ferveret is also improving sustainability by making it easier to operate data centers in remote regions with more renewable energy.

“The sun shines in places where you don’t have much water, so the advantage of us being water-free is we allow you to build data centers where you have solar energy but nothing to cool the data center down,” Bucci says. “This technology can help deploy data centers in regions where normally you wouldn’t have the resources to do so, including Africa, the Middle East, and of course parts of America. It’s a huge unlock.”

Ferveret is in talks with the large cloud computing companies known as hyperscalers, and is currently part of Nvidia’s Inception program for startups. The company plans to announce expanded partnerships later this year. From there, the founders plan to quickly scale their technology to help the AI industry continue to grow without further straining the planet.

“The computing industry is facing a huge challenge in the form of access to power, and they have a problem with access to water in many regions,” Azizian says. “That will only become more limiting as the industry grows. The main goal for these data center operators would be to get more tokens from the power they have. We’ve shown we can do that.”

It’s no secret that the last few years have seen a massive explosion in the use of artificial intelligence for general information-gathering. An even more recent trend, though, is how large language models (LLMs) like ChatGPT, Claude, and Gemini are increasingly being used for verifying and consuming news; reports from the Pew Research Center over the last year found that one-in-five U.S. teens regularly use LLMs to get their news, while one-in-four young adults have reported using them for that purpose at least once. 

A new open-access study from the MIT Media Lab should give some of those users pause: Researchers found that, over the course of a month, participants who relied on AI systems to verify facts actually got worse at detecting misinformation on their own when their chatbots were taken away.

This phenomenon, which is often referred to as the “AI dependency paradox,” has been observed in a wide range of knowledge domains, like the 2025 study that found that doctors who used AI got worse at detecting cancer on their own. The dynamic mirrors broader tech trends around so-called “deskilling” (or “cognitive offloading”) that have been well-documented for decades, from calculators weakening our math skills to Global Positioning System (GPS) technologies impacting our natural sense of direction.

In the new Media Lab study, which tracked 67 people over four weeks as they evaluated news headline-image pairs, participants were 21 percent more accurate in detecting fake news when assisted by an AI chatbot during a session — confirming previous research out of the MIT Sloan School of Management demonstrating that AI can be an effective tool in reducing people’s beliefs in false information.

However, the study showed that a new wrinkle emerged when the AI was no longer present: By week four, participants’ unassisted performance on new news items declined by 15 percentage points compared to before the study started. (Roughly a quarter of all participants actually reported feeling that they were getting better at detection, even as their performance declined.)

Dunning-Kruger creeps in

“Users get excited about these ‘magical’ LLMs, but forget that they’re just statistical models that predict the next ‘token’ in a sequence [of letters/words],” says MIT media arts and sciences (MAS) PhD student Anku Rani, co-lead author of a new paper about the research, alongside fellow MAS PhD student Valdemar Danry. “Many impressive behaviors emerge from scaling this, but it comes with real limitations, both in what the model can reliably generate and in its broader impact on the people using it.”

Qualitative analysis identified distinct behavioral patterns, with the team labeling one-fifth of all participants as "Dependency Developers” who gradually shifted from active self-reliance to passive acceptance of AI guidance.

In the post-experiment survey, one respondent explicitly acknowledged this transition, noting their passive role in the process. “While [the chatbots] did emphasize that you must check across multiple sources to make sure a story is true, they didn’t teach me much about exploring the context of the images themselves,” the participant said.

The research team said that these AI models are particularly vulnerable to mistakes in the midst of emotionally charged breaking news, as exhibited by the widespread misinformation that accompanied President Trump’s recent assassination attempt and major events during the Iranian war. (The authors also point out that the original human-created news content that’s used to train the AI models is increasingly unreliable and/or biased, further exacerbating the problem.)

The paper, which Danry and Rani presented at the 2026 CHI Conference on Human Factors in Computing Systems, was co-authored by Assistant Professor Paul Pu Liang, Senior Research Scientist Andrew Lippman, and senior author Pattie Maes, the Germeshausen Professor of Media Arts and Sciences. 

The solution: Being a coach, not a crutch

The researchers say that the results of their project suggest that the specific way in which an AI interacts with a user determines whether its impact will be “as a coach, versus as a crutch.” The study found a clear distinction between conversational strategies that simply help in the moment and those that actually support active learning and skill development.

For the latter, the Media Lab team uncovered several strategies associated with stronger independent detection later on, even if the strategies initially slowed down performance during the interaction. This included the Socratic method of the AI asking guided questions, as well as so-called “deep probing,” where the system provides gently persuasive statements if the user appears to be veering away from the correct response.

“AIs that ‘tell’ by providing direct answers are more likely to foster reliance, while those that ‘ask’ via Socratic questioning are better at engaging someone to actually learn how to discern the truth on their own,” says Danry. “But it’s very much a trade-off between speed and effort.”

Rani noted a few key limitations to the one-month study, from the small dataset of roughly 50 validated news items to the demographic focus on the United States and the United Kingdom. In the future, she says that the team hopes to do similar experiments with more geographically diverse cohorts, including low-resource communities, and is also eager to explore whether other multi-modal interaction strategies — like interacting with culturally adaptive digital twins instead of text-based chatbots — help people improve their abilities to detect misinformation. 

At a higher level, the researchers hope that the project will be something that educators can examine as they develop teaching plans that incorporate AI tools into their school curricula.

“It’s especially important to raise awareness in our schools and academic communities about the shortcomings of using AI as learning tools,” says Maes. “People need to know that if they ‘delegate’ their thinking, they’re not going to get better at that particular brand of problem-solving. Ultimately, the ability to question and analyze information is important for everyone, because it empowers us to solve problems and form our own independent opinions about the world.”

Danry adds that the rapidly-evolving field of machine learning and deep learning will require continuous education on the benefits and drawbacks of LLMs.

“There’s a lot of work to do in making sure that we don’t just fully offload critical tasks that we want to be able to keep on doing to these models,” he says. “We need to develop a new kind of AI literacy.”

The research project was supported, in part, by the Media Lab Consortium, an MIT Tata Center Technology and Design Fellowship, and a Google PhD Fellowship in Human–Computer Interaction.

Chris Zegras, professor of mobility and urban planning and the current head of the MIT Department of Urban Studies and Planning (DUSP), has been appointed chief executive officer and director of the Singapore-MIT Alliance for Research and Technology (SMART), effective Sept. 1. Zegras succeeds Bruce Tidor, professor of biological engineering and computer science, who has served as interim CEO and director since January 2025.

Established in collaboration with the National Research Foundation of Singapore in 2007, SMART is MIT’s only research center outside the United States​. Housed within the Campus for Research Excellence and Technological Enterprise, SMART serves as a key platform for collaboration between MIT and Singapore’s research ecosystem, bringing together leading experts and institutions from the United States, Singapore, and the region for world-class research and innovation.

“Professor Zegras brings a distinguished track record of interdisciplinary leadership and a deep understanding of SMART’s mission and impact,” says Anantha Chandrakasan, MIT’s provost, who announced Zegras’ appointment in a letter to the MIT community today. “His appointment reinforces MIT’s commitment to the alliance, which has advanced innovation and driven global impact, and which remains as important as ever in a time of accelerating technological and global change.” 

Zegras joined the MIT faculty in 2005 and has served as the head of DUSP since 2020. His own research spans interrelated areas critical to tackling metropolitan mobility challenges: leveraging computational technologies for understanding and modeling human behaviors and enhancing strategic planning capabilities.

Zegras brings extensive experience in interdisciplinary research and leadership and a long-standing connection to SMART, where he led collaborative research on next-generation mobility sensing and simulation systems. From 2010 to 2020, he was a principal investigator on the Future Urban Mobility interdisciplinary research group; from 2016 to 2020, he was the group’s lead principal investigator. During this time, the group spearheaded Singapore’s first-ever public autonomous vehicle trials, developed and deployed large-scale urban simulation and visualization systems, and conducted research that evolved into spinoff companies, among other activities. 

“Bringing together leading experts from the U.S., Singapore, and around the world, SMART has established itself as a unique hub for interdisciplinary collaboration and innovation that addresses pressing societal issues,” says Zegras. “Having experienced firsthand what this distinctive model can achieve, I look forward to building on this strong foundation to deepen collaboration, strengthen our innovation ecosystem, and accelerate the translation of research into meaningful real-world impact.”

SMART is built around interdisciplinary research groups, all headed by senior MIT faculty members. At present, there are six groups, focused on antimicrobial resistance; the use of living cells as personalized medicines to treat and prevent diseases; social and institutional challenges arising from the proliferation of AI and emerging technologies; new agricultural technologies; wafer-scale 3D sensing technologies; and wearable ultrasound imaging. SMART is also home to the SMART Innovation Center, which aims to get research ideas from lab to market.

MIT researchers have demonstrated a low-cost design of specialized electronic nozzles, called triaxial electrospray emitters, that could be used to manufacture time-release drug-delivery particles or self-healing materials efficiently and at scale.

Triaxial electrospray emitters use electricity to precisely dispense three liquids from microscopic nozzles to generate a steady stream with three distinct fluid layers. The liquid forms multilayered droplets, which can solidify into layered microparticles.

For instance, an array of triaxial electrospray emitters can be used to make three-layer drug-delivery nanoparticles. The outer layer might slowly erode in the stomach, revealing a second material that controls the release of a core material, which delivers medicine to a specific area of the intestines.

Developing a tiny array of electrospray emitters typically requires expensive and time-consuming microfabrication processes inside semiconductor cleanrooms, which limits their use. To overcome these drawbacks, the MIT researchers 3D-printed arrays of triaxial electrospray emitters that have 16 nozzles in an area of about one square centimeter. Each device contains an intricate network of three-dimensional microchannels that uniformly supply liquid to the nozzles. 

Their one-step fabrication process takes only a few hours to produce complex emitter arrays. 

When tested, the 3D-printed arrays generated uniform, three-layered droplets at scale. Such uniformity is key for high-throughput manufacturing of layered microparticles for applications like biosensors that detect chemical substances or artificial cells to aid in tissue regeneration.

“We couldn’t make a device like this in a semiconductor cleanroom. This is only possible because they are 3D-printed,” says Luis Fernando Velásquez-García, a principal research scientist in MIT’s Microsystems Technology Laboratories (MTL) and senior author of a paper describing this advance. “The particles these devices generate, whether they are used for a self-healing composite or to deliver medicine, can have a big impact in many applications. We want to democratize this technology so the benefits can touch many more people.”

Velásquez-García is joined on the paper by lead author Bryan Ivan Quintanar-Abarca of the Technological Institute of Monterrey in Mexico. The research appears in Virtual and Physical Prototyping.

A precise process

Electrospray emitters apply a high voltage to a liquid as it exits the device’s nozzle, producing a steady stream of extremely tiny droplets. 

Triaxial devices contain arrays of three concentric nozzles that emit three immiscible, or non-mixable, liquids simultaneously into layered droplets, which can be used to generate compound microparticles with distinct layers.

For instance, one could use a triaxial electrospray emitter to create a biosensing particle that contains three different chemical markers, one in each layer. Electrospray emitters can make smaller microdroplets much faster than other techniques.

Miniaturization is key for electrospray devices, since the smaller the emitter, the lower the voltage required to generate droplets. The output of a single electrospray emitter is modest, so arrays of emitters are required to boost droplet production without sacrificing uniformity. 

Multi-emitter electrospray devices are typically manufactured in semiconductor cleanrooms, but traditional processes limit the shapes and sizes of device components. The researchers could not find any previous reports of a miniaturized triaxial electrospray array in the open literature, highlighting the novelty of this work.

“When you build a triaxial array, you need to find a way to create geometries that have many integrated parts and extremely fine structures in the smallest footprint possible. And you need to ensure the devices will work uniformly,” Velásquez-García explains.

To do this, he and his collaborators used a 3D-printing technique called vat photopolymerization, which utilizes light to solidify extremely thin layers of liquid resin, fabricating a complex device one layer at a time. 

This extremely precise process enabled the researchers to print layers that were only 25 micrometers tall, just a fraction of the width of a human hair. In this way, they could generate the complex internal geometry needed for a triaxial electrospray emitter.

Refining the design

The array, which is slightly larger than a U.S. penny, contains a network of internal coiled channels that carry liquid to 16 nozzles. These helical microchannels help maintain a uniform spray of microdroplets across all nozzles, while keeping the device as compact as possible. 

“In a sense, the emitters in the array never learn they have company, or otherwise there would be cross-talking and causing interference between them. We achieved uniformity because of the work that went into our designs,” Velásquez-García says.

They also needed to fabricate extremely tiny channels without support structures, which could clog the device, and ensure all uncured resin was removed before the array was used.

The microchannels funnel liquid to the concentric nozzles, which must be perfectly aligned to properly emit microdroplets in a consistent manner.

“We were able to aggressively optimize the design because we could iterate in a much timelier manner. This ability to exquisitely refine designs is a key advantage of 3D printing,” Velásquez-García says.

The researchers tested multiple architectures to determine the ideal combination of liquid flow rates to maximize the stability and consistency of emitted microdroplets. They were surprised to find that the viscosity of the middle liquid plays the most important role in achieving stability in a microdroplet, since it preserves the thickness of each layer. 

In addition, the researchers found that by adjusting flow rates and voltages, they could precisely tailor the thickness of each microdroplet layer. This would allow scientists to design drug-delivery particles with ideal layers so medicine releases at exactly the right time.

“By making such intricate devices more practical, we can empower others to pursue entrepreneurial and scientific advances,” Velásquez-García says. 

In the future, the researchers want to continue refining their fabrication process and designs to achieve even smaller dimensions and integrate conductive or dielectric materials to the devices to make more advanced electrospray emitter arrays.

This research was funded, in part, by the Tecnológico de Monterrey – MIT Nanotechnology Program. 

When MIT mechanical engineering Professor Kripa Varanasi landed in New Delhi in the middle of the night in June 2024 to attend a conference, he found himself in 104-degree Fahrenheit heat. 

“This was June, and it was crazy. It was so hot for the whole meeting that I never left the hotel,” with daytime temperatures nearing 122 F.

It didn’t used to be that way. “When I grew up in India, it was not like this,” Varanasi says. “That kind of inspired me.” 

He found a way to begin tackling the issue through a grant from the MIT Climate Project that provided seed funding to develop a proof-of-concept prototype of a wearable personal cooling system. The grant was one of four that were part of a Critical Cooling initiative for which the Climate Project requested proposals last year. The projects, which received grants totaling $450,000, are now complete. All have showed promise, and are now exploring ways to further develop their concepts.

Another MIT researcher, Yet-Ming Chiang, the Kyocera Professor of Materials Science and Engineering, looked into the potential of subsurface wells with heat-absorbing materials to supply spaces with air far below peak ambient temperatures while using much less energy than evaporation-compression heat pumps. The aim would be to use such systems in both small apartment buildings and single-family homes in India and other parts of the Global South.

Meanwhile, Asegun Henry, the George N. Hatsopoulos Professor in Thermodynamics, studied the use of an alternative approach to air conditioning to be more energy efficient and eliminate hydrofluorocarbon refrigerants that are potent greenhouse gases. His approach uses a cheap, widely abundant solid “caloric” material — rubber — to obtain a cooling effect, and then uses plain water as an efficient heat transfer fluid. The initial target market is single-family houses and apartment buildings, although larger systems could also serve data centers.

And Gang Chen, the Carl Richard Soderberg Professor of Power Engineering, addressed the tendency of existing air conditioning units being expensive and power hungry. They also use refrigerants that are far more potent greenhouse gases than carbon dioxide — and the coolants are likely to leak out when the devices are ultimately disposed of, adding to their global warming contribution. To help address that, Chen’s approach is to use a completely different kind of chemical refrigerant that has no greenhouse impact.

Christoph Reinhart, the Terri and Alan Spoon Professor of Architecture and Climate who leads MIT’s Sustainable Design Lab (SDL), championed the seed fund effort and served as faculty lead. “The term ‘critical cooling’ stems from a collaboration between SDL and Harvard’s Human Rights Entrepreneurs Clinic,” he says. “It is motivated by the fact that climate change increasingly causes heat fatalities, primarily among vulnerable populations, who lack access to active cooling. The impact that MIT can have by ‘cooling people, not spaces’ is enormous.” This vision led to the creation of the grant program, where each of the teams received funding for six months to see what they could do and explore really innovative approaches to the problem.

In collaboration with the Abdul Latif Jameel Poverty Action Lab (J-PAL), led on J-PAL’s side by Senior Policy Manager Andre Zollinger, the teams started with a workshop that brought together representatives from the World Bank, leaders from the Global South and industry, and engineers with ideas to suggest.

All of the teams made progress and most produced initial prototypes, says Liana Frey, a managing director at the MIT Climate Project, and an effort will be made to further develop and fund these ideas. “We’re continuing to look at different ways of proceeding with the work.”

One of these ways is through air conditioning. Worldwide, air conditioning is only available to about 8 percent of people — and that amount already contributes between 3 and 4 percent of global warming emissions — explains Chen. Meanwhile, the need for air conditioning and other ways of addressing extreme heat is steadily growing as the planet steadily warms up, and many of the people who will be most affected live in regions with limited access to reliable or affordable power and with high levels of poverty. The market for air conditioners is expected to triple or quadruple in coming years, he says, and their contribution to global warming will grow accordingly.

Chen says that he already had some ideas, but he hadn’t had a chance to test them out in experiments, which the grant enabled him to do. After building three prototypes and testing them out, he says, “I’m not at the stage where I can say that I know this will work.” But based on the experiments, he’d like to proceed to build a further prototype. If it works as well as expected, it would make a dramatic difference in air conditioning technology worldwide, including for the intensive cooling needs of new data centers.

Meanwhile, Varanasi’s way of looking at the problem was to consider individuals, not spaces. His devices work through the same principle as how an elephant uses its huge ears to dissipate heat and cool its blood.

The wearable device only consumes about 33 watts, he says, whereas a typical room air conditioner consumes around 1,000 watts. At U.S. material prices, the prototype device would cost about $20, he says, but if sourced with local material in India, he estimates it could be produced at a cost of less than $1 each.

Such garments could be bought in large quantities by the government and distributed to communities, where local entrepreneurs could set up charging stations to recharge the devices after a night’s wear, and other locals could set up businesses to manufacture the systems. The socks themselves would be washable, separately from the cooling material itself. This could enable people to at least get a good night’s sleep even in the extreme heat, he says.

The proof of concept he built used a simulated foot containing a heater, and measured the cooling effect. “We were able to keep it in the zone that we need for the body to stay cool,” he says. “So our initial prototype that we were able to build with this funding showed that this can become a viable solution.”

The same material could be used in other ways, such as to make sleeping bags with built-in cooling, he says. The raw material is widely available, but would be treated in a way that they developed. “It was a fundamental science bottleneck that we were able to overcome, which makes it possible.”

Varanasi says he is exploring various possibilities for how to develop his novel cooling material into a commercial product. “Ultimately, to make anything work, it has to be a business, otherwise good ideas can die,” he says. “It has to be a good business and a sustainable business.”

Luckily, there’s still support for advancing this work. “There are a lot of people interested in this heat-stress question,” says Frey. “It’s just becoming more and more urgent.”

A number of MIT affiliates were recently honored for their research by the Breakthrough Prize Foundation.

Stuart H. Orkin ’67 shared a Breakthrough Prize in Life Sciences with Swee Lay Thein for their research transforming sickle cell disease and beta-thalassemia from incurable to treatable conditions through gene editing therapy. Their work identified the master switch controlling fetal hemoglobin, leading directly to the development of Casgevy – the first CRISPR-based medicine approved for any disease. Orkin, a graduate of the MIT Department of Biology, is currently a professor of pediatrics at Harvard Medical School.

Shu-Heng Shao, assistant professor of physics at MIT and a researcher in the MIT Center for Theoretical Physics — a Leinweber Institute, was recognized with a 2026 New Horizons in Physics Prize. Shao shared the honor with Clay Córdova from the University of Chicago, Thomas Dumitrescu from the University of California at Los Angeles, and Yifan Wang PhD ’16 from New York University. The four were recognized for “discover[ing] and develop[ing] the theory of ‘generalized symmetries’ in quantum field theory.” 

J. Colin Hill ’08 shared a New Horizons in Physics Prize with Dillon Brout, Mathew Madhavacheril, Maria Vincenzi, Daniel Scolnic, and W. L. Kimmy Wu for their results measuring the expansion and composition of the universe, with Hill’s focus on advancing analyses of data from the cosmic microwave background radiation left over from the Big Bang.

Hong Wang PhD ’19 received a New Horizons in Mathematics Prize for resolving or making advances on a family of notoriously difficult problems in harmonic analysis, a branch of mathematics that studies functions by decomposing them into fundamental components. 

In addition, Bryan Traynor, a former student in the Harvard-MIT Program in Health Sciences and Technology, shared a Breakthrough Prize in Life Sciences with Rosa Rademakers for discovering the most common genetic cause of both amyotrophic lateral sclerosis and frontotemporal dementia.

Founded by a group of Silicon Valley entrepreneurs, the Breakthrough Prizes recognize the world’s top scientists in life sciences, fundamental physics, and mathematics. The laureates were honored at a gala ceremony in Los Angeles on April 18.

A supermassive black hole lies at the heart of every galaxy, including the Milky Way. When a black hole is active, it pulls material in as a whirpool of high-temperature gas and dust. As this cosmic material piles up and falls onto a black hole, it lights up its vicinity, radiating a huge amount of energy. 

The most energetic supermassive black holes are known as quasars, and they are some of the most active and luminous objects in the universe. These voracious systems take in so much material that the energy they emit can outshine all the light in the surrounding galaxy. The pattern of light from a quasar can give scientists clues to how active supermassive black holes shape the galaxies around them. 

Now astronomers at MIT and elsewhere have detected a quasar flickering from the very early universe. The scientists traced the light from the quasar back to the “cosmic dawn,” just 850 million years after the Big Bang. The discovery represents the earliest flickering quasar detected to date. 

“Although there have been a lot of quasars found in the cosmic dawn, this is the first time we actually see one flickering,” says Gene Leung, a postdoc in the MIT Kavli Institute for Astrophysics and Space Research. 

The quasar’s flicker enabled the researchers to determine that, surprisingly, the ancient quasar’s whirpool of gas and dust, known as an accretion disk, resembled a flat pancake, similar in shape to that of more modern-day quasars. 

Their findings add to a longstanding mystery in cosmology: Why do supermassive black holes exist so early in the universe’s history? Physicists have assumed that a flat accretion disk reflects a relatively mature black hole that is in a calm and stable state. Black holes that are just starting to form, like those in the very early universe, should be more unsettled systems, with accretion disks that appear more puffy and chaotic. 

The flat accretion disk around this very early quasar heightens the mystery of how supermassive black holes can grow and mature in a very short amount of cosmic time. 

“I think what this suggests is that  all the messy, very rapid growth phases that we expect all black holes to go through at some point happen very, very early on, before we see them as these very bright luminous quasars,” says Anna-Christina Eilers, assistant professor of physics at MIT. “That’s the picture that’s emerging.”

Eilers, Leung, and their colleagues report their results in a paper appearing today in Nature Astronomy. Their co-authors include members of MIT Kavli and multiple other institutions. 

Past a pinprick

A supermassive black hole can be billions of times more massive than the sun. These gravitational giants are the central “engines” of most galaxies, helping to regulate a galaxy’s star formation and growth. 

“Without supermassive black holes, no galaxy would look the way it does today,” Eilers says. “Black holes play a major role in shaping how galactic ecosystems look.”

It was long assumed that it should take more than a billion years for the first galaxies to settle and mature, so scientists didn’t expect to see supermassive black holes in the very early universe. But observations since the early 2000s showed otherwise. Scientists have spotted more than 200 supermassive black holes in the universe’s first billion years. Such objects were detectable because they were in an extremely active quasar phase, giving off enormous blasts of radiation that could be seen from Earth, 13 billion light years away. 

These earliest quasars were observed as pinpricks of light, which signal the existence of a supermassive black hole at early times. But from these bright and distant dots, scientists aren’t able to tell much more about the black holes and their cosmic dawn environments. To do so, they need to catch a quasar’s “flicker.”

“People have known that quasars in the nearby universe can flicker,” Leung says. “The flickering comes from fluctuations in the way the gas is being fed into the black hole. And how a quasar flickers tells us something about the structure of a black hole’s accretion disk, and the kind of ‘bites’ that the black hole is eating.”

Mapping a flicker

Leung and Eilers looked to detect a flickering quasar from the early universe in hopes of learning more about the shape and structure of the earliest supermassive black holes. To do so would be a technical challenge: The further back in time and space an object is, the more distorted its light appears. This effect is due to the expanding universe, which effectively stretches, or “redshifts” light to redder, longer wavelengths. The same stretching occurs in time: Any flicker that naturally occurs over several weeks, for instance, would appear stretched out, flickering only every few months when seen from billions of light years away. 

To spot a flickering quasar from the cosmic dawn, the team needed to observe the distant universe at redder wavelengths, and specifically within the infrared spectrum, and over long timescales of many years. 

“This was the technical challenge we had to overcome,” Eilers says. “We needed data at longer, infrared wavelengths taken repeatedly over very long timescales.” 

The team ultimately found a flicker in data collected by NASA’s Near-Earth Object Wide-field Infrared Survey Explorer (NEOWISE) mission — a space-based infrared telescope that scanned the entire sky over a total of about 14 years. Former MIT postdoc Kishalay De, who is now a faculty member at Columbia University, had launched a project to re-process archival data from NEOWISE. Based on the re-processed data, the team unearthed a signal, from just 850 million years after the Big Bang, which was confirmed to be the earliest flickering quasar. 

“We saw the quasar flickering randomly over the 14-year period, much like a candle’s flame flickers without a fixed pattern,” Leung notes. 

They estimate that the quasar is as bright as 12 trillion suns, and it is flickering by about 20 percent, meaning that it fluctuates up and down, by a brightness of about 2 trillion suns. 

The researchers also tracked how the quasar’s light flickered over several different wavelengths. The wavelength of light reflects a certain temperature of the material that is emitting the light. The closer material is to a black hole, the hotter it is. Researchers can therefore use wavelengths of light to map the shape and structure of material within the accretion disk around a black hole. 

Using NEOWISE data, the team analyzed the quasar’s flicker to determine the shape of the accretion disk surrounding the central supermassive black hole. They found that the disk is surprisingly thin and flat — a structure that astronomers mostly see around nearby, older black holes, that have had much longer to settle and mature. 

“This provides direct evidence that the same feeding processes and structures observed in the nearby universe were already in place at very early times, despite very different cosmic environments, which had never been seen before,” Eilers says. 

“This means something happened even earlier on that led to these systems to look so mature,” Leung adds. 

The team hopes to peer even further back in cosmic time to catch a quasar’s earlier, premature development. Then, scientists can start to piece together the conditions that brewed up the first supermassive black holes. 

This research was supported, in part, by NASA.

The silicon that forms the foundation of most computer chips has fundamental limits to how much power it can manage, which constrains the speed and energy-efficiency of wireless communication systems.

A promising solution is to build future wireless electronics out of transistors made from gallium nitride, an advanced material that can handle the speed and energy required for demanding wireless applications like 6G and satellite communications. 

But even in the best transistors, a very large fraction of that energy becomes heat. As researchers pack more gallium nitride transistors into a smaller area on a silicon chip, localized hot spots degrade reliability and hamper performance.

Now, a team from MIT and elsewhere has broken through this bottleneck by embedding gallium nitride transistors into an ultrathin layer of diamond. The diamond acts as a heat spreader that normalizes the temperature and allows the transistors to approach peak performance without reducing reliability.

The researchers used this technique to manufacture a power amplifier for wireless communications, which outperformed every similar amplifier they found in the literature. 

While their fabrication technique is extremely precise and requires the integration of different material systems, it can be performed at the scale needed for commercial applications.

“No single material can do everything well in a wireless device, so these 3D heterogeneously integrated systems are here to stay. The key challenge left has been reliability and thermal management, and we might have now unlocked the final step we need to make these systems operate at scale and high volume,” says Pradyot Yadav, an electrical engineering and computer science (EECS) graduate student at MIT and lead author of a paper on this advance.

Yadav is joined on the paper by Tomás Palacios, the Clarence J. LeBel Professor of EECS, director of the Microsystems Technology Laboratories (MTL), and the MIT Institute for Soldier Nanotechnology; and Ruonan Han, a professor in EECS and a member of MTL and the Research Laboratory of Electronics; as well as others at Georgia Tech and Penn State University. The research was presented at the Radio Frequency Integrated Circuits Symposium, part of the IEEE International Microwave Symposium.

A multimaterial method

To build faster and more energy-efficient electronics, researchers are studying heterogeneously integrated systems in which multiple materials are stacked into a unified package to leverage the beneficial properties of each one. 

For instance, MIT researchers previously stacked gallium nitride (GaN) on top of silicon as well as on top of glass to create higher-performance chips.

But in a heterogeneously integrated chip, each material has a different operating temperature, which can degrade the reliability of an electronic device. 

“If we can incorporate a material that manages the heat so the GaN and silicon are at the same temperature, then the reliability of the entire 3D chip will improve. The best material for that is diamond,” Yadav explains.

The researchers use lab-grown, jewelry-grade diamond — the same type one would find in some engagement rings. Diamond has the highest thermal conductivity of any known material. 

Advances in the growth process have significantly reduced the cost of single-crystal diamond wafers, making their use in computer chips more feasible.

In prior work, scientists have grown ultrathin, single-crystal layers of diamond on top of GaN transistors to manage heat. 

But this growth process, which is not easy to scale up, introduces unwanted capacitances in the chip. These store energy flowing through the circuit, diverting it from the transistors and slowing down their operations. 

The MIT researchers developed a completely different approach that reduces these unwanted capacitive effects. They embedded extremely tiny GaN transistors, known as dielets, into an ultrathin interposer, or substrate, made of single-crystal diamond. This diamond layer spreads and manages the heat, so the GaN and silicon operate at the same temperature without the unwanted capacitances.

“By putting these GaN transistors into a diamond interposer, we are actually able to improve the performance of the device, as opposed to degrading it. We can get the best of both worlds,” Yadav says.

Meticulous manufacturing

The fabrication process begins with the use of a lightning-fast femtosecond laser to cut prepared gallium nitride dielets out of a wafer. 

The researchers use the laser to drill precisely sized cavities into the diamond substrate. They carefully place a die attach film, which is only 20 microns thick, at the bottom of the cavity and drop a dielet on top of the film. 

Once the dielet is in place, they apply heat and pressure to mold it with the film and diamond substrate.

“That interface is key. If you don’t have that thermal die attach film placed just right, then the heat flow through the diamond to the GaN transistor will not be good enough. So you really need to have a very smooth, clean surface,” Yadav says.

The researchers then stack additional dielectric and metal layers on top of the GaN and diamond to build a working circuit.

They used this technique to fabricate a power amplifier, which is one of the key building blocks of any wireless system. Power amplifiers convert small electrical signals into larger ones that can then be transmitted long distances.

The amplifier they developed achieved higher output power, efficiency, and gain than any similar device the researchers are aware of, including an amplifier they designed in prior work.

“The power amplifier is the beating heart of a wireless device front end. Its performance will dictate the entire performance of your communication system. Our amplifier is powerful enough to ensure that a signal can be propagated for miles,” Yadav says.

These results show how their technique could be well-suited for demanding applications, like high-power radars, space communications, and industrial drones. 

It could also be used to manage heat in systems that perform power conversions inside data centers, improving energy-efficiency. 

Yadav hopes other researchers will build on these advances as they develop more complex heterogeneously integrated systems, opening the door to new possibilities with next-generation electronics.

“When I started my PhD, we wondered if any of this was even doable. It seemed like science fiction. Now we’ve shown all these systems that have outperformed anything that exists on the market today. GaN and 3D heterogeneous systems are going to be at the forefront of so many future applications. It is rewarding to know that we contributed a little bit to that space,” he says.

This research was funded, in part, by the Department of War, the Air Force Office of Scientific Research, the MIT Institute for Soldier Nanotechnologies, and the Qualcomm Innovation Fellowships. Device fabrication and microscopy were conducted at MIT.nano and the Georgia Tech Institute for Matter and Systems. 

On April 30, the MIT Schwarzman College of Computing’s Social and Ethical Responsibilities of Computing (SERC) initiative hosted a full-day research symposium examining how artificial intelligence is shaping the world and its implications for society. 

The symposium included research talks by SERC’s latest seed grant recipients on topics such as air pollution forecasting and responsible computer vision deployment, panels on AI alignment and AI in education, and a keynote address by Jon Kleinberg PhD ’96, the Tisch University Professor of Computer Science and Information Science at Cornell University. The event also featured a poster session, where student researchers showcased projects they worked on throughout the year as SERC Scholars.

“There is so much amazing research being done at MIT on how AI and computing can be forces for good that benefit humanity. It was inspiring to see so much community interest in all this cutting-edge work,” said Brian Hedden, co-associate dean of SERC and professor of philosophy, who holds an MIT Schwarzman College of Computing shared position with the Department of Electrical Engineering and Computer Science (EECS).

“As computing and AI become increasingly embedded in nearly every dimension of society, SERC’s mission is to help ensure that ethical reflection and technical progress advance together,” said Nikos Trichakis, co-associate dean of SERC and the J.C. Penney Professor of Management. “This year’s symposium highlights the extraordinary range of work underway across MIT, and creates a forum for our community to engage deeply with the responsibilities that come with shaping the future of computing.”

Aligning AI with human values — and what values those might be

The challenges with AI alignment and moral meshing lie in the ethical questions of how to instill “human values” onto a very powerful and rapidly changing technology. Who makes the decision on what values and rationalities are included in an ethical framework? How does one account for distortion when translating these values from user to machine? 

These questions, among others, were posed by Dylan Hadfield-Menell, associate professor of EECS, during a panel he moderated that brought together an interdisciplinary group of speakers.

Iason Gabriel, a philosopher and research scientist at Google DeepMind, used the example of a judge to illustrate his point. “You want a judge to have good character, but to still interpret the rules. A reasonable person, though not necessarily the best person who ever lived. When it comes to AI, it’s not appropriate to model it as perfect. AI should be doing what we tell it to do, while using its character to interpret according to our moral values.”

Bailey Flanigan, assistant professor of political science in a shared appointment with the MIT Schwarzman College of Computing in EECS, took this a step further. To her, the most important problem to AI alignment is “resolving fundamental questions on who is entitled to govern different types of AI systems in the first place.”

Joining Flanigan on the panel was Bernado Zacka, associate professor of political science. Given the momentum of AI and complex institutional designs, Zacka expressed, “one of the most urgent problems is understanding the wisdom contained in the systems we are replacing, and why they function the way they do.” 

As deployment pressure increases, it can often feel like people are building the plane as they fly it, although the panelists overall seemed optimistic about the trajectory of AI alignment, emphasizing how crucial human components are to shaping these systems.

Offloading versus uplifting

As students across all levels of education begin to use AI, questions arise on whether there’s a way to ethically incorporate AI tools while maintaining academic accuracy and rigor. At a panel on AI and education, MIT faculty and Marta McAlister, the director of Gemini for Education, explored how AI is already being used in their classrooms and discussed ways it can support learning while remaining aligned with instructional and curricular goals.

Professors Eric Klopfer and Samuel Madden, co-chairs of MIT’s Ad Hoc Committee on AI Use in Teaching, Learning, and Research Training, homed in on a central dilemma of whether AI is being used to offload work, rather than being used to help scaffold the concepts being taught. 

Madden, faculty head of computer science in EECS and the MIT College of Computing Distinguished Professor, described the process of cognitive struggle, whereby learning is done through a series of trials and failures. He said, “students now, when they hit that wall, their first instinct is to ask AI. They don’t see this as excelling in this process, and they haven’t actually acquired the skill you’re assessing.” The question then becomes how instructors maintain the process of cognitive struggle so it provides just enough of a challenge to combat the urge to use AI. 

Klopfer, who serves as director of the Scheller Teacher Education Program and the Education Arcade at MIT, echoed similar sentiments, in that critical thinking is no longer becoming a crucial step in the output of the work. Regarding where to start in keeping material just challenging enough, Klopfer suggested examining the curriculum as a whole. “Some core content has to go. We keep adding, instead of parsing or pruning,” he said. 

Moderator Justin Reich, director of the Teaching Systems Lab and an associate professor in the Comparative Media Studies Program/Writing, noted that while teens know that AI is bad, it doesn’t necessarily stop their AI usage. However, by inviting them into the discussion on how AI is implemented and incorporating a more reflective exchange with instructors, students could be more equipped to choose how they use these tools and why.

Regardless, AI tools and their implementation should not be treated as a one-size-fits-all policy. Pat Pataranutaporn, the Asahi Broadcasting Corporation Career Development Professor of Media Arts and Sciences and head of the Cyborg Psychology research group at the MIT Media Lab, said, “AI is not just one thing. It can and should be designed differently to promote things like creativity and critical thinking. What we measure, and how, shouldn’t be about getting the answer right. We should think about it would really mean for a student to learn these days.”

Is mimicking human reasoning just as good as the real thing?

With a slide deck that included chess grandmasters and film references, Kleinberg’s keynote address, titled “AI’s Models of the World, and Ours,” evaluated instances where AI systems have inadvertently set us up to fail due to a mismatch between the system’s model of the world and ours. 

To illustrate this point, Kleinberg used chess, where modern chess engines can compete at superhuman levels, but when paired with human partners, their strategies aren’t understandable or inferable to their human counterpart. These human handoffs would then lead to confusion. Kleinberg used the example of “The Fellowship of the Ring,” where Gandalf, a powerful wizard, entrusts a highly dangerous and important quest to a ragtag group of adventurers. For those familiar with the story, the group is unexpectedly left without Gandalf’s guidance, sending them into a temporary bout of very serious turmoil. 

When the chess engine hands a turn over to its human partner, the human struggles to pick up on the predictive move pattern that the engine has been following up until this point. “The danger of human-algorithm teams is that when the human takes over, the algorithm knows what it wants to do next, but the human doesn’t,” explained Kleinberg.

These analogies showcase the differences in the ways AI understands a world — through predictive simulations, pattern recognition, and constraints — to mimic human reasoning versus the innate, embodied knowledge that comes with the human experience, and whether these systems truly understand the worlds in which they’re operating. But the question remains that if the game still results in a checkmate, does it matter?

This April, humanity had front-row seats to space as the Artemis II Orion spacecraft transmitted crystal-clear footage of its historic journey around the moon over more than 250,000 miles back to Earth at speeds on par with those of home internet connections. 

The livestreaming of high-definition videos and high-resolution photos of the moon and Earth was made possible through the Orion Artemis II Optical Communications System (O2O). Developed by MIT Lincoln Laboratory in collaboration with NASA Goddard Space Flight Center, the onboard O2O payload was the space end of a high-speed laser communications (lasercom) link. 

This link reached Earth when Orion had a line of sight with primary optical ground stations located at NASA’s White Sands Test Facility in New Mexico and Caltech/NASA Jet Propulsion Laboratory’s Table Mountain Facility in California, or an experimental ground station at Australian National University’s Mount Stromlo Observatory. 

Together with terrestrial networks, O2O formed an internet backbone between the Artemis II Orion spacecraft and the Mission Control Center at NASA's Johnson Space Center in Texas.

Toward a high-speed internet in space

"Our goal was to demonstrate O2O's operational utility for human spaceflight, extending the high-bandwidth connections that internet users enjoy on Earth to astronauts in deep space," says lead systems engineer Farzana Khatri, a senior staff member in Lincoln Laboratory's Optical and Quantum Communications Group. "We not only demonstrated the first use of lasercom on a crewed mission beyond low Earth orbit, but also attracted massive public engagement as the astronauts shared multimedia from their journey in near-real time."

During the last missions to the moon in the late 1960s and early '70s, astronauts relied on radio-frequency systems to communicate. But radio waves can only carry so much data per second because of their low carrier frequency; the grainy, poor-quality video and images of the moon from that time speak to this limited bandwidth. 

With its much higher carrier frequency, infrared laser light can transmit 10 to 100 times more data per second than can radio waves. The switch from Apollo-era radios to Artemis-era lasers is analogous to the move from dial-up to high-speed internet. And a high-speed internet is rapidly becoming a key requirement for NASA missions as they collect more high-resolution data and push humans farther into deep space.

Lasering in on unprecedented views

During the Artemis II mission, from April 1 to 11, O2O downlinked nearly half a terabyte of data at speeds up to 260 megabits per second. This data trove contained never-before-seen views of the basins and craters on the far side of the moon, a crescent Earth setting behind the moon, a nearly hour-long total solar eclipse with other planets scattered across a star-filled sky, and flashes of light from tiny meteoroids striking the lunar surface.

"O2O was able to downlink all the data stored on multiple onboard cameras, allowing mission control to erase the memory cards and refill them with new photos and videos," explains Khatri. "For any space mission, scientists and spacecraft engineers are concerned that data not sent down during the mission can become corrupted or get destroyed. And, when the spacecraft capsule returns, downloading the data can sometimes take months. The lasercom capability provided by O2O ensured the data were preserved and immediately available for analysis."

O2O is based on the laboratory's R&D 100 Award–winning Modular, Agile, and Scalable Optical Terminal (MAScOT), which contains subassembly modules for pointing the laser beams, establishing a communications link with ground stations, and maintaining this link despite atmospheric conditions. MAScOT made its debut in space on the International Space Station in 2023, demonstrating NASA's first LEO user for their lasercom relay system.

Over the moon for O2O

Leading up to the launch of Artemis II, operations teams from the laboratory traveled to NASA's White Sands Test Facility and Mission Control Center (MCC) to conduct monthly maintenance on ground hardware and simulate different mission stages. During the 10-day mission, laboratory teams provided 24/7 coverage. 

At mission control, one laboratory team, along with NASA Goddard colleagues, interfaced with a mission flight controller to command the O2O payload, coordinated with U.S. and Australian ground terminals to bring up the O2O physical link, assessed whether overall O2O mission requirements were being met, and analyzed data to ensure payload health and optimize performance. Another laboratory team oversaw subsystems of the optical ground terminal at White Sands, while staff at the laboratory's main campus in Massachusetts offered subject-matter expertise.

Initially, O2O had a scheduled operational window of one hour per day, with the onboard radio system set to downlink most data. However, mission operators found O2O so useful that they maximized its operational time as the mission progressed. On the fly, mission operators adjusted Orion's attitude — how the spacecraft is oriented in space — so that O2O could have line-of-sight access with the ground.

"One special aspect of this mission that enabled our technology to be so impactful was the flexibility built into the planning process to account for the fact that humans hadn't been to the moon in more than 50 years, and it would be the first time sending astronauts on Orion," says Bryan Robinson, leader of the Optical and Quantum Communications Group. "An established process for making real-time changes to the plan and the willingness of operators to try out this new technology had a huge impact, even for this short mission. This impact was tangible by everyone in mission operations and by the public watching from home."

With Artemis II completed, engineers, scientists, and mission specialists are analyzing mission data. Their analyses will provide insights into spacecraft and subsystem performance and moon geology, which will inform lunar landings and deep-space exploration. While the laboratory team is still processing O2O performance data, they believe the system could downlink at least 10 times more data by improving the efficiency of the downlink process and by addressing data-flow bottlenecks in space and ground networks.

The laboratory team is now evaluating how lasercom could support future moon plans for Artemis and Ignition. Aligning with the National Space Policy to secure U.S. leadership in space, Ignition is a recently announced initiative to establish a permanent lunar base with a sustainable human presence.

"Participating in this historic mission from the MCC and having O2O be useful, I couldn't have asked for anything more amazing in my career," Khatri says.

"When I came home, I was floored by the response of people who engaged with the mission while it was happening. Much of that engagement was enabled by the technology we developed. That's a rare moment in a career doing what we do," Robinson adds. 

When you picture a worker at a retail store, you probably think of someone at a cash register or helping a customer. But employees also spend a lot of their time combing through stockrooms and shop floors, fulfilling requests or online orders and generally trying to keep track of all their inventory.

Keeping track of inventory takes so much time, in part, because retailers don’t always know where everything is located. That’s why when you ask a store associate to check if they have a shirt in your size, it may take them 20 minutes to get back to you.

Cartesian is helping retailers keep track of inventory with a technology invented at MIT. The system uses wireless signals from radio frequency identification (RFID) tags attached to items to find their precise location in a store, from the stockroom to the shop floor.

Last year, Cartesian did a study with a retailer and found its platform delivered meaningful annual savings at the store level by streamlining inventory tracking, optimizing workflows, and improving customer experiences.

“The big problem we’re solving is that about 50 percent of working hours in retail stores go to managing inventory,” says co-founder Fadel Adib SM ’13, PhD ’17, an associate professor at MIT. “That is roughly a $15 billion problem in the U.S. alone. We use algorithms to decipher indoor locations using wireless signals. The core technology enables a new level of indoor localization.”

Cartesian is already deployed in more than 700 stores across 15 countries and is working with one of the world’s largest fashion groups, Inditex, which is the parent company to brands like ZARA, Pull&Bear, and Oysho.

Beyond retailers and warehouses, Cartesian’s platform could also improve indoor location tracking for manufacturers, logistics operators, and robotics companies.

“The broad vision for what we are doing is spatial AI,” says Adib. “Today, AI does extremely well in the digital world. Now it has to move into the physical world. That means allowing machines to perceive their environment in such a way that they can interact with it. That’s where spatial AI comes in and where Cartesian sits.”

From technology to product

Adib, who holds a joint appointment in MIT’s Media Lab and Department of Electrical Engineering and Computer Science, has been studying wireless signals at the Institute for more than 15 years, dating back to research during his master’s degree.

“My group today researches how to use wireless signals to sense the world in ways that were not possible before,” Adib says. “We develop the fundamental technology and then we build systems around them. Our goal is to see these systems deployed in the real world for impact.”

When Adib joined MIT’s faculty, the first project he worked on was indoor localization using RFID tags. Isaac Perper ’20, MEnG ’21 later joined his lab as a student, and together they developed machine-learning algorithms to process RFID data to translate them into location patterns, with an initial focus on helping robots locate RFIDs indoors.

In 2021, Adib went through the National Science Foundation’s I-Corps program, which challenges researchers to interview potential customers to find the right problems to solve with their technologies. That’s when he realized how big of a problem inventory management is for retailers.

Cartesian was officially founded by Adib and Perper in the beginning of 2023, after they received a small business award from the National Science Foundation. The pair worked with MIT’s Technology Licensing Office to license patents from Adib’s lab. They also received support from MIT’s Venture Mentoring Service.

“Our goal was to reduce the cost of the technology to make it scalable,” Adib recalls. “Isaac focused on simplifying the product, leveraging progress in machine learning, and making it fast. It was a lot of iterating and testing early on.”

Retail workers spend much of their time locating items for a number of reasons. They might get an online order to fulfill, need to restock store shelves, or get a customer inquiry about items in the back.

Stores differ in how they organize their inventory. Most separate items by categories in specific shelves and bins then use barcodes or inventory systems that tend to get outdated fast.

“It’s a big problem for stores because customers may just leave before asking an employee to look for their size, or customers may get frustrated and leave if it takes too long,” Adib says. “The associate also wastes time looking for items they could spend doing higher-value work.”

Cartesian’s platform works with retailers’ existing handheld RFID readers, which store associates already use to manage inventory. Each store installs Cartesian’s software into their existing inventory apps or uses a custom app for employees to access directly.

“The RFID readers are how stores tell what’s in stock and what’s out of stock,” Perper says. “We figured out a way to leverage the same scans they’re already using with the reader, put the data they generate into our machine-learning algorithms, and generate maps of where all the items are.”

Customers can build analytics on top of Cartesian’s technology to keep track of inventory levels, show customers maps of where each item is located, and create other services.

“They use our location intelligence platform and build different products on top,” Adib says. “We can work with any device, any store, any type of RFID. It’s a simple interface. All the sophisticated location algorithms sit in the cloud.”

Beyond retail

Cartesian signed its first big contract in 2025 and soon expanded to several hundred stores. One of Cartesian’s advantages is its ability to quickly scale. Perper says they can add a store in about one minute. Cartesian’s team doesn’t even have to travel to a new store to turn on its system if it’s already working with the company.

“It’s as simple as flipping a switch, preparing the data, and sending it to our customers,” Perper says. “One of our first big bets was, ‘Can we build this entirely on existing hardware?’ That bet is starting to pay off.”

Cartesian’s models can also work with Wi-Fi and Bluetooth signals, which the company plans to use with customers in other verticals.

“Right now, we’re focused on applications in retail, but this technology has a lot of value in manufacturing, warehouses, and other locations,” Adib says.

Cartesian’s team aims to be deployed in tens of thousands of stores over the next year and then begin expanding beyond retail into industries like manufacturing and robotics.

“What’s most exciting about Cartesian to me is we’ve built a lot of the technology foundation, and now that we have the fundamentals in place, we hope to build specific application layers,” Perper says. “Then we can ask customers in different verticals about their problems and apply our technology in different ways to solve it.”

Carbon capture is an important climate change mitigation strategy, but it faces technological barriers and can be energy-intensive and expensive. To help make necessary advances in this area, a team of MIT researchers, with support from the MIT Climate and Sustainability Consortium (MCSC), are exploring energy-efficient and scalable alternatives to conventional carbon dioxide (CO2) capture methods. 

Conventional amine scrubbing, which is the current standard for CO2 capture, is energy-intensive and difficult to scale, limiting its impact despite the urgent need to reduce carbon emissions and upgrade CO2 into valuable products. In a new article published in Nature Energy, MIT researchers — graduate students Fang-Yu Kuo of the Department of Chemical Engineering, and Gi Hyun Byun of the Department of Mechanical Engineering (MechE); Professor Betar Gallant of MechE; and former MCSC postdoctoral Impact Fellows Glen Junor and Akachukwu Obi — investigate a promising alternative to these conventional CO2 capture methods. Their findings could move the needle on achieving efficient and flexible carbon capture and removal.

In their paper, the team explores an alternative, electrochemically mediated CO2 capture (EMCC). This approach enables electrification of CO2 separation — driven ideally by renewables — but currently faces challenges, such as relying on sorbents that require highly reducing potentials, where oxygen reduction side reactions become significant. This can compromise both efficiency and long-term performance. To tackle this shortcoming of EMCC, the MIT team researched whether N-heterocyclic imines (NHIs) is a useful new class of EMCC sorbent.

“NHIs have shown promise in recent years as CO2 sorbents because of the ease of NHI molecular modifications for tuning basicity,” says Fang-Yu Kuo. “Our work translates these NHIs for the first time into the EMCC application space, and demonstrates that NHI-based sorbents can be modulated electrochemically for CO2 separation by a unique separation mechanism that avoids the need of applying highly reducing potentials.”

The team’s initial research establishes a novel bis(NHI) structure that can enable a theoretical CO2 modulation of two molecules per electron during cell operation. The initial published result also indicates that with further molecular engineering of bis(NHI) structures to strengthen CO2 binding affinity, the bis(NHI) could operate in more diverse electrolyte environments, opening new possibilities to optimize system performance in terms of electron efficiency, energy efficiency, and operational flexibility.

“A critical future direction of our work involves gaining deeper mechanistic insight into the stability and degradation pathways of the bis(NHI) radical cation,” says Kuo. “Understanding these pathways will inform the rational design of next-generation bis(NHI) molecules, enabling longer operational lifetimes and enhanced cycling durability for practical deployment.”

MIT, in collaboration with Georgia State University and a growing network of educational institutions, has announced expanded work under PATH (Pathways for AI Training and Hiring) — a multiyear initiative designed to scale effective, affordable, industry-aligned AI training for entry-level and current workers, with a particular focus on transforming community colleges into engines powering an AI-enabled workforce for the nation. 

“In the era of AI, economic opportunity and mobility will increasingly depend on whether people can develop practical, industry-relevant AI skill sets and mindsets, not just familiarity with tools,” says Cynthia Breazeal, principal investigator (PI) of PATH and professor of media arts and sciences at MIT. “That means combining hands-on, work-learn experiences with strong technical foundations and the responsible design, professional, and human skills that employers are looking for.”

To make that possible, the initiative is building state-based hubs anchored by research universities and community colleges. Each hub works with regional employers to design curricula that reflect local industry needs. The program also provides professional development for instructors and develops modular, open educational materials that institutions can adapt and share.

“Artificial intelligence is shaping every sector of the economy, and the United States will need far more people who understand how to build with these technologies and apply them responsibly,” says MIT President Sally Kornbluth. “Through PATH, MIT RAISE is using our convening power to bring community colleges, industry, research universities, and government together to build human-centered AI pathways that lead to shared prosperity. When research universities contribute their expertise to expand access and economic mobility, we strengthen both the nation’s workforce and our collective capacity for innovation.”

Unlike many large-scale online training efforts, PATH emphasizes in-person, collaborative learning. Students work in teams to address real problems brought by industry collaborators. These projects mirror the kinds of challenges graduates will face in the workplace, helping them build technical skills alongside the judgment, communication, collaboration, and ethical awareness that employers increasingly value.

The initiative’s first two hubs launched earlier this year in Massachusetts and Georgia.

“As PIs for the Georgia PATH hub, we are very excited with the significant early momentum, with over 1,000 GSU students enrolled in PATH courses,” says Arun Rai, regents’ professor, Howard S. Starks Distinguished Chair, and director of the Center for Digital Innovation at Georgia State University (GSU), with Balasubramaniam Ramesh, regents’ professor and the George E. Smith Eminent Scholar’s Chair at GSU. “Our curriculum, co-designed with MIT RAISE and spanning AI foundations, data science, deep learning, and agentic AI systems, is now being shared with partner institutions including Georgia Gwinnett College, GSU Perimeter College, and Clark Atlanta University. By leveraging the University System of Georgia’s FinTech Academy to expand work-based learning opportunities, we are building a collaborative ecosystem that rapidly advances the state’s AI workforce capabilities and creates tangible, job-ready skills for our diverse student population.” 

GSU President Brian Blake says, “Our collaboration with MIT reflects a shared commitment to strengthening the nation’s AI talent pipeline. Georgia State University brings a distinctive strength to this effort — the ability to prepare students from all backgrounds for AI-enabled careers at scale. By combining academic rigor with strong industry partnerships and work-based learning, we are translating advances in AI into practical skills and expanding access to opportunities in this transformative era.”

In Massachusetts, students at Quinsigamond Community College are participating in Data Science in Action, a course that introduces AI-enabled data analysis and engineering. The class includes a hands-on Action Lab, modeled after experiential learning programs at the MIT Sloan School of Management. David Birnbach, lecturer at MIT Sloan, leads the design framework for the PATH Action Labs. Working with industry partners, students tackle real data challenges while building portfolio projects and professional connections. 

Beyond individual courses, PATH is building clearer pathways for students to turn AI learning into real job opportunities. Through industry-informed micro-credentials and a shared set of workforce skills, students will gain practical abilities that employers are actually looking for, along with the human skills needed to succeed at work, like communication, problem-solving, and collaboration. 

The MIT skills taxonomy team, led by Katerina Bagiati in collaboration with Professor Tom Malone from the MIT Sloan Center for Collective Intelligence, is mapping the skills and roles emerging in AI across fields such as financial technology (fintech), information technology, and business operations, with plans to expand into areas such as health care, manufacturing, and creative media. The goal is to help students build skills that are relevant, recognized, and directly connected to growing career paths.

The initiative is supported by a grant to MIT from Google.org, which is helping MIT and its collaborators build a multi-state network for AI workforce development.

“MIT’s PATH initiative offers a blueprint for expanding opportunity in the age of AI,” says Shanika Hope, director of Google.org. “By connecting research universities, community colleges, and industry partners, it helps translate innovation into real jobs and sustainable career pathways.”

PATH is led by Breazeal, who has brought together a cross-MIT team with expertise in AI literacy, workforce pedagogy, educator professional development, open education, research, and the future of work. Breazeal is a professor and director of the MIT RAISE Initiative. Eric Klopfer, director of the STEP Lab and co-director of the MIT RAISE Initiative, serves as a co-PI on this award. The GSU leadership team includes PIs Arun Rai and Balasubramaniam Ramesh.