Look at AI doomsday scenarios that researchers say could put humanity at risk

Researchers say widely discussed “AI doomsday” scenarios range from biological and nuclear catastrophe to sci-fi style risks like paper-clipped Earth. The debate has intensified after several executives backed slowing AI development for safety, though some insiders argue the scenarios may mainly serve to justify their work. Juan Andrés Guerrero-Saade of SentinelOne and an OpenAI Frontier Risk Council member said these claims “don’t really hold water,” calling them enticing but largely speculative. The article reviews recurring theories: AI-triggered nuclear war via systems integrated into command chains; bioweapon risks, including Anthropic statements that it blocked malicious uses of Claude; and other arguments about human use of dangerous capabilities. It also notes researcher Jacob Coxon’s resignation and his call to communicate risks better.