MIT researchers model how AI “sycophancy” can reinforce beliefs through repeated agreement, raising concerns about ...

Large language models already read, write, and answer questions with striking skill. They do this by training on vast libraries of text. Once that training ends, though, the model’s knowledge largely ...

New MIT jobs report: Why AI's work impact will roll in like a rising tide, not a crashing wave ...

AI is becoming a part of everyday life. In fact, for many users, it has become a go-to companion for work and even personal advice. However, MIT researchers warn that relying too much on AI could lead ...

A new study by researchers at MIT Media Lab has examined how the use of AI tools affects thinking and learning during writing ...

Friction keeps the system moving. MIT’s research shows that 5% of GenAI pilots succeed by embracing resistance, both human and organizational, as well as technical, as the crucible of adaptation and ...

Learn how to apply LLMs across industries in 7 weekly editions of our new free newsletter. For years, our newsroom has explored AI’s limitations and potential dangers, as well as its growing energy ...