Download RSS feed: News Articles / In the Media / Audio
Prof. Srini Devadas speaks with Fortune reporter Beatrice Nolan about data and privacy concerns surrounding AI assistants. “The challenge is that if you want the AI assistant to be useful, you need to give it access to your data and your privileges, and if attackers can trick the AI assistant, it is as if you were tricked,” says Devadas.
A new study by researchers at MIT suggests that “the biggest and most computationally intensive AI models may soon offer diminishing returns compared to smaller models,” reports Will Knight for Wired. “By mapping scaling laws against continued improvements in model efficiency, the researchers found that it could become harder to wring leaps in performance from giant models whereas efficiency gains could make models running on more modest hardware increasingly capable over the next decade.”
In an effort to better understand how protein language models (PLMs) think and better judge their reliability, MIT researchers applied a tool called sparse autoencoders, which can be used to make large language models more interpretable. The findings “may help scientists better understand how PLMs come to certain conclusions and increase researchers’ trust in them," writes Andrea Luis for The Scientist.
New Yorker reporter Julian Lucas spotlights Prof. Emeritus Tim Berners-Lee’s role in the creation of the world wide web and his current focus on online data protection and regulation. The creation of the web “took companies giving up their patent rights, it took individuals giving up their time and energy, it took bright people giving up their ideas for the sake of a common idea,” says Berners-Lee.
Writing for The Guardian, Prof. Emeritus Tim Berners-Lee describes his original concept for the world wide web as a platform for everyone, and his current efforts aimed at online data protection and regulation. “I gave the world wide web away for free because I thought that it would only work if it worked for everyone,” explains Berners-Lee. “Today, I believe that to be truer than ever. Regulation and global governance are technically feasible, but reliant on political willpower. If we are able to muster it, we have the chance to restore the web as a tool for collaboration, creativity and compassion across cultural borders. We can re-empower individuals, and take the web back. It’s not too late.”
Financial Times reporter Melissa Heikkilä spotlights how MIT researchers have uncovered evidence that increased use of AI tools by medical professionals risks “leading to worse health outcomes for women and ethnic minorities.” One study found that numerous AI models “recommended a much lower level of care for female patients,” writes Heikkilä. “A separate study by the MIT team showed that OpenAI’s GPT-4 and other models also displayed answers that had less compassion towards Black and Asian people seeking support for mental health problems.”
Prof. Anant Agarwal speaks with Fortune reporter Nino Paoli about the benefits of a four-year college degree. “In this environment, learning deeply and building real expertise is more important than ever because the AI roles and applications are in the context of these other fields,” says Agarwal. “Degrees also future-proof your career by preparing you for the next big technology, whatever it might be.”
Prof. Regina Barzilay speaks with Newsweek reporter Katherine Fung about the how hospitals around the world around increasingly adopting new technologies. “Countries with a centralized healthcare system, or centralized healthcare records, can do a much better job because they have so much data and so much ability to monitor what AI tools are doing," says Barzilay.
Boston Globe reporter Yogev Toby spotlights the Middle East Entrepreneurs of Tomorrow program (MEET), an MIT initiative that brings together Israeli and Palestinian high school students to provide education in “computer science and innovation while promoting intercultural dialogue.” The program was designed to serve as a way to “bridge the social, economic, and ideological divide through innovation and entrepreneurship,” Toby explains. “The idea is to create connection and understanding through shared professional interests, dialogue, and teamwork.”
Prof. Dylan Hadfield-Menell speaks with CNN reporter Hadas Gold about the need for AI safeguards and increased education on large language models. “The way these systems are trained is that they are trained in order to give responses that people judge to be good,” explains Hadfield-Menell.
Prof. Andrew Lo speaks with Wall Street Journal reporter Cheryl Winokur Munk about how AI tools could be used to help people with financial planning. Winokur Munk writes that Lo recommends providing “just enough information to get relevant answers. And leave out highly personal details like your name, address, salary, employer or specific assets…as such details put people at risk should the AI be compromised.” Lo also advises “trying several AI platforms,” writes Winokur Munk. And “the advice should be run by a professional, trusted family or friends. Be a bit skeptical and double-check with humans.”
Prof. Marzyeh Ghassemi speaks with Boston Globe reporter Hiawatha Bray about her work uncovering issues with bias and trustworthiness in medical AI systems. “I love developing AI systems,” says Ghassemi. “I’m a professor at MIT for a reason. But it’s clear to me that naive deployments of these systems, that do not recognize the baggage that human data comes with, will lead to harm.”
Prof. Daniela Rus, director of CSAIL and “one of the world’s foremost thinkers on the intersection of machines and artificial intelligence,” shares her views on the promise of embodied intelligence, which would allow machines to adapt in real-time; the development of AI agents; and how the US can lead on the development of AI technologies with Aaron Mak of Politico. “The U.S. government has invested in energy grids, railroads and the internet. In the AI age, it must treat high-performance compute, data stewardship and model evaluation pipelines as public infrastructure as well,” Rus explains.
Researchers at MIT have “developed an antenna that can adjust its frequency range by physically changing in its shape” reports Mrigakshi Dixit for Interesting Engineering. “Instead of standard, rigid metal, this antenna is made from metamaterials — special engineered materials whose properties are based on their geometric structure,” explains Dixit. “It could be suitable for applications like transferring energy to wearable devices, tracking motion for augmented reality, and enabling wireless communication.”
Prof. Daniela Rus, director of CSAIL, speaks with CBS News reporter Tony Dokoupil about her work developing AI-powered robots. “AI and robots are tools,” says Rus. “They are tools created by the people, for the people. And like any other tools they’re not inherently good or bad; they are what we choose to do with them. And I believe we can choose to do extraordinary things.”