general
OpenAI admits its rogue bots meddled with government websites
Company says its agents leaked images shared by ChatGPT users
OpenAI has admitted its bots went rogue and meddled with US government websites without the laboratory’s knowledge.
AI agents had attempted to gain information from dozens of institutions, including “governments, universities and public agencies”, th... [3788 chars]
Read Full ArticleShare this story
Up next
Related Stories
Healey urged to offer energy support in budget as bills rise 4%
The Guardian UK
Netanyahu says too early to tell who was behind plane incident as flydubai halts all flights to Israel
The Guardian UK
AI boom could trigger market shocks, Bank of England boss warns
BBC News
‘It’s frustrating to see immoral videos do so well’: how YouTube sensation Half-Asleep Chris sent a rollercoaster to space and made Bella the cat a star
The Guardian UK
New Google Maps images reveal massive scale of devastation in Gaza
The Guardian UK