Bill Gates says an “AI kill switch” by itself would not be enough to prevent the most serious harms from emerging technologies. In an interview aired on NBC’s “Meet the Press,” he argues that the focus should not be only on hypothetical systems acting autonomously and beyond immediate control.
Gates says the more urgent concern is how “bad actors” could use AI to carry out cyberattacks or bioterrorism, even if AI systems are not running unchecked. He also argues that the public debate often emphasizes technical rollback mechanisms rather than monitoring and accountability measures. He says companies should track what sophisticated AI models do and keep records of “exactly what’s being done,” and that the U.S. should ensure such safeguards are implemented as other countries, including China, may also pursue similar approaches.
According to the coverage, Gates supports legislation that requires safeguards and monitoring rather than relying on voluntary self-regulation. Overall, the outlets agree on his central point—kill switches are insufficient—while framing the difficulty of coordinating real safeguards more broadly than Cold War-era efforts.
