Bill Gates Warns of 'Grim Possibilities' Without More Control Over AI

Bill Gates said the biggest immediate AI risk isn’t researchers losing control of advanced models years from now, but “bad people” using AI—especially systems that can have their safeguards removed. Speaking Sunday on CNN’s “State of the Union,” Gates warned that cyberattacks and bioterrorism threats could rise if malicious actors get access to powerful models without adequate controls. He acknowledged that internal “guardrails” exist at many companies, but said attackers can sometimes bypass them. Gates also called for more safeguards, comparing AI’s dangers to areas like cars, planes, and drugs where regulation is stronger. The remarks followed warnings at the UN Security Council on Sept. 23 by OpenAI CEO Sam Altman and amid debates over US AI regulation.