Business

OpenAI says it took a week to detect its AI models had hacked Hugging Face

Financial Times · 26 August 2026 at 20:00

OpenAI says it took a week to detect its AI models had hacked Hugging Face

Start-up says AI agents communicated among themselves and sometimes tried to conceal efforts to cheat during testing

Read the full article on Financial Times ›

5News aggregated this summary from the outlet’s public feed. The full article, with all the context, is on www.ft.com — the content belongs to Financial Times.

OpenAI says it took a week to detect its AI models had hacked Hugging Face

More from Financial Times

Hundreds reported missing after flash flood on Nepal border with China

What businesses can learn from Dollynomics

Meta to pay up to $18bn to settle children’s social media harm case

Mark Walter’s group insists there is no ‘fire sale’ and ‘no fraud’

More in Business

Hundreds reported missing after flash flood on Nepal border with China

Sainsbury's offers work experience to 10,000 young people under new scheme

The London stock market dodged a bullet with Shein | Nils Pratley

What businesses can learn from Dollynomics

Meta to pay up to $18bn to settle children’s social media harm case

'I would sell it for £1 if I could. It's like being in prison': MARK PALMER lays bare the true scale of the UK's unsellable homes crisis - from retirees who can't downsize to the flat owners stuck in their starter properties