Computer Science and Artificial Intelligence Laboratory (CSAIL)

Prof. David Sontag, Monica Agrawal PhD '23, Luke Murray SM '22, and Divya Gopinath '19, MEng '20 co-founded Layer Health - an AI healthcare startup that is applying large language models (LLMs) to help clinicians with medical chart reviews and data abstraction, reports Seth Joseph for Forbes. “The same chart review problem we’re solving with our clinical registry module is faced by clinicians at the point of care,” says Sontag. “For example, one of our next modules will focus on real-time clinical decision support to help automate clinical care pathways, leading to more reliable, high-quality care."

Prof. Erik Demaine and his colleagues have solved a longstanding mathematical puzzle called “Dudeney’s dissection,” which involves dissecting an equilateral triangle into the smallest number of pieces that could be rearranged into a square, reports Lyndie Chiou for Scientific American. “Each of those pieces could have arbitrarily many edges to it, and the coordinates of those cuts start at arbitrary points,” says Demaine, of what makes the puzzle so challenging. “You have these continuous parameters where there’s lots and lots of infinities of possible choices that makes it so annoyingly hard. You can’t just brute-force it with a computer.”

Venti Technologies – a company co-founded by MIT researchers and alumni – has been named one of the most innovative companies in the Asia-Pacific region for Fast Company’s 2025 roundup of top companies, reports Katerina Barton. The company focuses “on autonomous technologies for industrial use—specifically in low-speed environments like ports, airports, and warehouses,” explains Barton. “The company’s suite of special-purpose algorithms is designed to optimize cargo container transportation and works with a wide range of vehicles, allowing the AI-enabled technology to move varying weight loads and distances through complex spaces and changing routes.” 

[Originally in Spanish] MIT researchers have developed a new technique to educate robots by increasing human input, reports Uriel Bederman for TN Tecno.  “We can’t expect non-technical people to collect data and fine-tune a neural network model," explains graduate student Felix Yanwei Wang. "Consumers will expect the robot to work right out of the box, and if it doesn’t, they’ll want an intuitive way to customize it. That’s the challenge we’re addressing in this work."

Researchers from MIT and elsewhere estimate that “while 36% of U.S. private sector jobs were technically ‘exposed’ to automation through computer vision… it would only make economic sense for firms to pursue automation for 8% of all private sector jobs—just a quarter of those jobs labelled ‘exposed,’” report François Candelon, David Zuluaga Martínez and Etienne Cavin for Fortune. 

Writing for Forbes, Paula Schneider, CEO of Susan G. Komen, highlights Prof. Regina Barzilay’s research using AI to detect breast cancer. “Using her own mammograms in her research at MIT, Dr. Barzilay demonstrated how AI could have detected her breast cancer much earlier, potentially improving her prognosis,” writes Schneider. “Studies show that incorporating AI into mammogram analysis boosts cancer detection rates by 20%, without increasing false positives. This is a significant leap forward, as early detection is key to a better chance at positive outcomes and survival.” 

Prof. Sarah Millholland, Prof. Christian Wolf, Prof. Emil Verner, Prof. Darcy McRose, Prof. Marzyeh Ghassemi, Prof. Mohsen Ghaffari and Prof. Ariel Furst have received the 2025 Sloan Research Fellowship for “being among the most promising scientific researchers currently working in their fields,” reports Michael T. Nietzel for Forbes. “Sloan Research Fellows are chosen in seven scientific and technical fields—chemistry, computer science, Earth system science, economics, mathematics, neuroscience, and physics,” explains Nietzel. 

Writing for Wired, Prof. Daniela Rus, director of CSAIL, highlights the future of “physical intelligence, a new form of intelligent machine that can understand dynamic environments, cope with unpredictability, and make decisions in real time.” Rus writes: “Unlike the models used by standard AI, physical intelligence is rooted in physics; in understanding the fundamental principles of the real world, such as cause-and-effect.”

Researchers from MIT and elsewhere have found that “AI doesn’t even understand itself,” reports Peter Coy for The New York Times. The researchers “asked AI models to explain how they were thinking about problems as they worked through them,” writes Coy. “The models were pretty bad at introspection.” 

Liquid AI, an MIT startup, is developing technology that “holds the same promise of writing, analyzing, and creating content as its rivals while using far less computing power,” reports Aaron Pressman for The Boston Globe. 

Prof. David Autor has been named a Senior Fellow in the Schmidt Sciences AI2050 Fellows program, and Profs. Sara Beery, Gabriele Farina, Marzyeh Ghassemi, and Yoon Kim have been named Early Career AI2050 Fellows, reports Michael T. Nietzel for Forbes. The AI2050 fellowships provide funding and resources, while challenging “researchers to imagine the year 2050, where AI has been extremely beneficial and to conduct research that helps society realize its most beneficial impacts,” explains Nietzel. 

Prof. Daniela Rus, director of CSAIL, speaks with NBC Boston reporter Colton Bradford about her work developing a new AI system aimed at making grocery shopping easier, more personalized and more efficient. “I think there is an important synergy between what people can do and what machines can do,” says Rus. “You can think of it as machines have speed, but people have wisdom. Machines can lift heavy things, but people can reason about what to do with those heavy things.” 

Using a new technique developed to examine the risks of multimodal large language models used in robots, MIT researchers were able to have a “simulated robot arm do unsafe things like knocking items off a table or throwing them by describing actions in ways that the LLM did not recognize as harmful and reject,” writes Will Knight for Wired. “With LLMs a few wrong words don’t matter as much,” explains Prof. Pulkit Agrawal. “In robotics a few wrong actions can compound and result in task failure more easily.”

Prof. Armando Solar-Lezama speaks with New York Times reporter Sarah Kessler about the future of coding jobs, noting that AI systems still lack many essential skills. “When you’re talking about more foundational skills, knowing how to reason about a piece of code, knowing how to track down a bug across a large system, those are things that the current models really don’t know how to do,” says Solar-Lezama.

Prof. Daniela Rus, director of CSAIL, and Prof. Russ Tedrake speak with the Financial Times about how advances in AI have made it possible for robots to learn new skills and perform complex tasks. “All these cool things that we only dreamed of, we can now begin to realize,” says Rus. “Now we have to make sure that what we do with all these superpowers is good.”