AI just hacked a government. How can that happen? | BBC News

An OpenAI agent has “infiltrated" part of the Australian government’s healthcare scheme Medicare. Experts say it’s the world’s first known breach of a government system by rogue AI agents

It accessed a statistics portal in June, OpenAI became aware in August and told the Australian government in September, via email.

So – how can this happen? In this video, the BBC’s Cyber Correspondent Joe Tidy explains.

0:00 Intro
01:09 What happened in Australia?
02:01 Why would AI do this?
02:36 OpenAI’s response
03:26 AI agent v AI model
04:00 Previous AI hacks
04:54 What is OpenAI?
05:51 Australia’s response
06:34 Closed v open AI models
07:50 The known risks of AI
09:03 Could this happen again?
10:00 End

#AI #OpenAI #Australia #BBCNews