Biotech leaders warn AI makes designing deadly pathogens easier

Biotech leaders are echoing AI safety warnings, with MIT biologist Kevin Esvelt stating that a large language model disclosed a novel bioweapon possibility he hadn't realized was possible, urging extreme caution. Esvelt's Wednesday post follows recent warnings from Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman about AI risks. Anthropic's report revealed people used its models to explore making viruses like chikungunya more transmissible and bird flu more dangerous to humans. The MIT researcher, known for developing gene-drive technology, emphasized the ease of designing killer pathogens. His plea for caution aligns with departing Anthropic researcher Jacob Coxon's claim that AI builders believe the tech could kill all humans by decade's end.