AI Workers Scoff at Doomsday Warnings From Industry Peers

Employees at OpenAI, Meta and DeepMind are privately sceptical of viral warnings that unchecked AI development could kill people en masse, the BBC reports. Workers called the claims vague and joked about them, though many still acknowledge near-term risks such as jailbreaks and military adoption. More than 100 AI staff signed a letter backing independent safety evaluators in major labs.