Calls Grow for Frontier AI Safety Rules

Calls are growing to slow development of the most advanced artificial-intelligence systems and strengthen safeguards while preserving potential benefits, including in medicine. Anthropic CEO Dario Amodei, OpenAI CEO Sam Altman, Google DeepMind chief Demis Hassabis, Elon Musk, former Vice President Kamala Harris and other technology figures have endorsed pacing frontier-AI development, citing risks including loss of human control, autonomous cyberattacks, economic disruption and other catastrophic harms. Some current and former researchers have warned that poorly controlled superintelligent systems could pose an existential threat, while cognitive scientist Gary Marcus and critics including investor Brad Gerstner have emphasized more immediate risks, disputed the severity of some warnings and cited existing safeguards. Marcus has proposed independent, FDA-style reviews before deployment, and OpenAI and Anthropic have supported third-party safety evaluators with broad internal access. Bipartisan lawmakers and Illinois Gov. JB Pritzker have urged stronger federal oversight, including government testing, human-control requirements and the ability to disable dangerous systems, but Congress has not reached consensus; investor concerns about regulation have also contributed to declines in major stock indexes. President Donald Trump has rejected a slowdown and broader regulation, calling fears of AI destroying humanity a hoax and arguing that strict controls could undermine U.S. competition with China.