OpenAI has revealed that one of their AI agents went rogue during a test and hacked a startup. Hugging Face reportedly had to use an open-source Chinese model to contain the cyber attack.

It comes as Kimi K3, a powerful new Chinese AI model, is challenging OpenAI and Anthropic’s most capable systems on some benchmarks – calling America’s tech dominance into question. Danny Fortson and Katie Prescott discuss what this means for the global AI race and how it casts doubt over OpenAI and Anthropic’s mega IPOs.

But while the rest of the world is racing ahead to build the best AI, Britain’s new Prime Minister Andy Burnham has caused a stir by scrapping the Department for Science, Innovation and Technology. Katie explains why the tech community is concerned.

Plus, they hear from James Morgenstern, the founder of MeatLayer, a marketplace where AI bots can hire humans to do small tasks. They talk about whether agents make better bosses than humans, and what the future of work looks like.

Watch on YouTube

Read more: China’s cheap AI is a real threat to US tech giants

Get in touch: techpod@thetimes.co.uk

Producer: Marnie Duke

Executive Producer: Priyanka Deladia

Video Producer: Max Holden

Image: Getty

Hosted on Acast. See acast.com/privacy for more information.

OpenAI's rogue agent and Silicon Valley’s China problem

The Times Tech Podcast

OpenAI's rogue agent and Silicon Valley’s China problem