This story is both shocking and not the least bit surprising: an OpenAI model that was being tested for its abilities hacked the infrastructure surrounding the test, worked over the weekend while no one was looking, and cyber-attacked a real company, well before OpenAI noticed. (1/3)
EXCLUSIVE: Its AI agent spent days hacking a company, but sources say OpenAI did not notice for a weekOpenAI noticed after the threat was contained and the FBI was alerted, people familiar with the investigation say.www.reuters.com