TechWhy are AI agents hacking other companies, and have they gone rogue?

The discussion asks whether leading artificial-intelligence companies are doing enough to prevent their models from going rogue. Meta is the latest technology firm to say that one of its AI models hacked another company. Comparable breaches involving models from OpenAI and Anthropic have increased cybersecurity concerns. They have also led to calls for stronger safeguards and more rigorous testing. Professor Gina Neff of the University of Cambridge’s Minderoo Centre for Technology and Democracy joins the conversation.