AI researchers warn of doomsday scenarios including nuclear war, bioweapons, and internet takeover

Artificial intelligence researchers have long warned that AI could wipe out humanity through scenarios like nuclear war, bioweapons, or an internet takeover. These warnings have intensified as some executives call for slowing AI development. However, skeptics like Juan Andrés Guerrero-Saade dismiss them as sci-fi, while Anthropic researcher Jacob Coxon resigned over safety concerns. Anthropic recently blocked attempts to use its AI for bioweapon research, and its CEO warned of AI botnet attacks. The RAND Corp. report noted safeguards currently prevent AI from triggering nuclear war, but risks could grow if AI is integrated into decision chains.