AI experts warn of superintelligent machines and urge regulation amid US-China rivalry

In a moderated video debate, Oren Etzioni and Max Tegmark examined the accelerating pace of AI development, suggesting that recursive self-improvement could soon produce machines surpassing human research capabilities. Etzioni, once an optimist, now warns that many breakthroughs he expected years away are arriving faster than anticipated, and he stresses the danger of U.S. over-regulation handing the lead to China. Tegmark, formerly a pessimist, expressed renewed optimism that public concern is pushing AI into the mainstream, yet he also calls for a regulatory body akin to the FDA to enforce safety standards.

Both dismissed industry self-regulation as a “joke,” citing past failures in social media and tobacco. They concur that the safest approach is to keep AI systems compartmentalized, granting them only the permissions needed for specific tasks. While Etzioni fears a populist backlash could produce harmful policies, Tegmark argues that any nation developing superintelligence must act responsibly, regardless of trust in rivals.