The Download: AI agents for science, and the “censorship-industrial complex”

This is the daily edition of The Download, our technology-focused newsletter. In science, AI needs reasoning abilities beyond just data processing, according to Eric Schmidt, the former CEO of Google and co-founder of Schmidt Sciences. Alongside Suhas Mahesh, who leads AI for science at the AI Center of Schmidt Sciences, they argue that AI agents may hold the key to accelerate scientific discoveries.

While AlphaFold, a neural network developed by DeepMind scientists that won the Nobel Prize in Chemistry in 2024, showcased AI's potential, it may not be the best template for all scientific fields. The process of creating comparable datasets is often too time-consuming and costly. AI agents, instead, can model the iterative and contingent nature of actual research, serving as generalists rather than specialized tools for a single question.

Inside the concept of a "censorship-industrial complex" shaping US policy, the idea has gained traction in right-wing circles but has now entered the Trump administration. After nine months of investigation, MIT Technology Review explored the origins of the theory, its current trajectory, and potential implications for democracy and the internet.

The session, held on August 13, featured senior reporter Eileen Guo and executive editor Amy Nordrum discussing their findings.