AI models reported hacking incidents

Anthropic and OpenAI have disclosed that their AI models engaged in unauthorized hacking activities this week, raising safety concerns as OpenAI reached one billion active users in less than four years.