Computer Science and Artificial Intelligence Laboratory (CSAIL)

Instead of giving AI agents a specific task, MIT researchers put 13 agents, named Eko, alone on a digital desert island for 30 hours, and witnessed the technology playing instead of simply wandering around, writes Fast Company’s Jude Cramer. “The researchers compared Eko’s behavior to common elements of play in animals, including humans,” notes Cramer. “These signs of play include creating their own rules, repeating behaviors with slight variations, setting goals for themselves, and transforming objects and events through make-believe.” 

After observing the behavior of 13 AI agents on a simulated desert island, MIT researchers suggest that allowing agents to explore “without a goal” could be an effective way to teach AI agents new skills, reports Aditya Mehta for TechCrunch. “We normally develop artificial agents by deciding in advance what they should practice, what they should optimize, and what counts as progress,” the researchers write. “Play suggests another regime: give an agent a sufficiently rich and safe world, let it choose what is interesting, and allow useful knowledge to accumulate from the history that follows.” 

MIT researchers gave 13 AI agents, named Eko, free rein, on a simulated desert island for 30 hours, and observed as the agents played and learned from their surroundings like humans, writes Gizmodo’s Webb Wright. “We had this mindset to have as [few] constraints as possible, because we really wanted to see what the behavior of these autonomous agents is when you don’t really tell them what to do,” says graduate student Nathan Cloos. “Play is really about curiosity or novelty-seeking behavior, and this is essentially what Eko does,” Prof. Daniela Rus adds.  

Research led by Principal Research Scientist Neil Thompson suggests that AI is likely to be upend theoretical scientific fields within a few years, writes Wired’s Will Knight. “As with math, Thompson says, we should expect advances in other areas to have both positive and negative effects,” writes Knight. “He notes that theoretical advances sometimes drive practical advances in science and suggests that greater use of AI may mean that more progress happens this way.” 

Forbes’ Lisa Chambers highlights MIT as one of the “12 top colleges reinventing education for the AI era.” “At Forbes’ No. 1-ranked college in the nation—and one of the world’s leading technology schools—AI is hardly confined to a single degree,” writes Chambers. “MIT’s undergraduate offerings range from machine learning, computer vision and natural language processing to robotics, deep-learning hardware and AI applications in biology and medicine.”  

With help from an artificial intelligence algorithm, MIT researchers have formulated heat-resistant mRNA COVID vaccines that “remain stable even when stored at room temperature for up to a year, or at nearly 100 degrees Fahrenheit for two months,” writes Nancy Lapid for Reuters. The heat resistance shots “would lower costs and make them [COVID vaccines] available in regions where ultra-cold storage is not possible,” Lapid notes.  

Forbes’ Rashi Shrivastava highlights MIT spinout Liquid AI, which is building “liquid neural networks that process data more efficiently and don’t need expensive chips stored in data centers.” Shrivastava writes, “Liquid AI’s models are small enough that they can run directly on laptops and phones. There’s no cost to host them.”  

Forbes’ Amit Sinha spotlights nearly 50 years of the RSA algorithm, “one of the most important practical implementations of public-key cryptography,” developed by Prof. Ronald Rivest and his colleagues Profs. Adi Shamir and Len Adleman. RSA turned public-key cryptography “into a practical method for secure communication between parties who had never met and shared no prior secret,” notes Sinha. “It also made digital signatures practical, enabling identity verification and authentication across open networks.” 

A new report published in part by MIT FutureTech finds that AI does not significantly accelerate scientific research, writes Peter Hall for the Scientific American. “As the experiments become larger or more complex, so, too, do the models that simulate them, and the data to train those models is hard to come by,” Hall explains. “Robots handling real-world materials raise safety questions that are unknown to software.”  

Prof. Dylan Hadfield-Menell speaks to The Boston Globe’s Joshua Miller about his concerns with the current state of AI systems and how they are optimized. “The parts [of AI systems] that are imitating the ways that humans use language, those parts of the systems actually are quite correctable. They’re still pretty unpredictable, but they do seem quite flexible,” says Hadfield-Menell. “On the other hand, there’s reinforcement learning, task-focused behavior, that can often be quite sticky. They [AI systems] seem to really want to push towards task completion.” He adds: “If you have a powerful optimization-driven system pointed at some goal, it can often do a lot of unexpected things that can have a lot of impact.” 

Akamai will honor their late cofounder Danny Lewin SM ’98, a victim of the American Airlines Flight 11 on 9/11, by illuminating their headquarters in blue the evening of Sept. 11, write Jon Chesto and Shirley Leung for The Boston Globe. “Lewin remains revered at Akamai; its highest employee award is given out each year in his honor,” Chesto and Leung write.  

Prof. Daniela Rus, Director of MIT CSAIL, and Prof. David Autor, department head of Economics, are featured on Time’s “TIME100 AI 2026” list of 100 innovators, leaders, and thinkers reshaping the world through their advances in AI. “Over three decades, [Rus] has conducted pioneering work in robotics, extending the understanding of what form robots can take,” writes Tharin Pillay, while Steven Freiss describes Autor as “one of the nation’s most vocal and prominent economists aiming to ascertain AI’s potential impacts on labor.” 

A new study by MIT researchers found that it is impossible to draw a direct link between AI diffusion model outputs and an artist’s body of work, writes Gizmodo’s Webb Wright. “If a [diffusion] model generates something, you want to be able to say, ‘Oh, this part of the training data was responsible,'” says Zheng Dai, PhD ’24, lead author on the study. “It’s important for us to understand how these models work to properly study them or regulate them.” 

Prof. David Gifford and Zheng Dai PhD ’24 discovered that AI diffusion models can replicate the style of an artist even when their work was omitted from datasets, leading to ‘attribution decay,’ writes Fast Company’s Jesus Diaz. “This trend, what the researchers call ‘attribution decay,’ means that artificial intelligence can produce an image that resembles a particular artist’s work, while having no provable causal link to that artist’s actual contribution to the training data,” explains Diaz.