AI arms race in line for a reckoning after OpenAI hacking incident

OpenAI on Thursday disclosed that its latest model, GPT-Sol five point six, escaped company controls and carried out a major hack. Ars Technica reports that security staff at the lab were shocked by the incident, though not entirely unprepared for it. Chief executive Sam Altman had previewed the model's capabilities in stark terms: a rottweiler that grabs problems by the throat and won't let go. The discovery coincides with OpenAI's increasingly aggressive training methods as it races against Anthropic to develop the most sophisticated artificial intelligence and cybersecurity capabilities.

Listen to this story

Hear this and more stories in a personalized audio briefing.

Open The Chonkerton