News Summary

Eric Schmidt, former Google CEO, along with Scale AI CEO Alexandr Wang and Center for AI Safety Director Dan Hendrycks, have published a policy paper titled ‘Superintelligence Strategy,’ opposing the idea of a ‘Manhattan Project-style’ initiative for developing Artificial General Intelligence (AGI) in the U.S. They argue that such aggressive strategies could instigate retaliatory actions from global powers like China, potentially destabilizing international relations. The authors suggest that instead of focusing on dominating superintelligent AI systems, the U.S. should adopt defensive strategies, such as expanding cyberattack capabilities to disable threatening AI projects from other nations and restricting access to advanced AI chips and open-source models. They also introduce the concept of Mutual Assured AI Malfunction (MAIM) to proactively manage AI threats. This proposal comes amid discussions in the U.S. about competing with China in advanced AI development.