DarkGPT, militarized models, “unbridled” agents… On black markets, cybercriminals are snapping up new clandestine AI

Also read

Between fantasies and the risk of “Apocalypse”, the real threats of artificial intelligence under the scrutiny of French researchers

SURVEY - In recent months, alerts on the existential risks linked to the development of AI have multiplied. While Silicon Valley raises the specter of an uncontrollable machine, French researchers and experts are trying to separate the real danger from the apocalyptic story.

AI: Developers don't know how to stop models from setting their own goals, says former Anthropic researcher

Jacob Coxon, who slammed the door of Anthropic at the beginning of September, believing that AI “could kill us all by the end of the decade”, was interviewed this Monday in New York.