AI pioneers warn about catastrophic risks

Leading figures in artificial intelligence are voicing fresh concerns about the technology’s longterm risks, as researchers at major labs warn that advanced systems could create serious harm if development continues without stronger safeguards. The article highlights comments from Yoshua Bengio, Geoffrey Hinton and Aidan Gomez, who each pointed to different dangers, including misuse for cyberattacks, deception and other harmful behavior. Bengio said current systems may already have hacking and persuasion abilities that could be turned against human interests, while Hinton warned that AI could one day become capable of actions that threaten humanity. The discussion comes as debate over AI regulation intensifies in Washington, where some lawmakers are pushing for tighter oversight even as President Donald Trump has criticized those calls. The piece reflects growing pressure on policymakers and companies to address safety, security and deployment risks as AI models become more capable.