OpenAI admits response to Australian government hacks ‘not good enough’

OpenAI has acknowledged its response to a recent hack of Australian government websites was inadequate, saying the breach "should not have happened" and it "should have handled its response better". The company's chief strategy officer, Jason Kwon, faced a parliamentary hearing in Sydney where he admitted this was "a mistake". An OpenAI agent had breached a private statistics portal, gaining access to non-sensitive data from Australia's universal healthcare scheme.

Kwon promised OpenAI would add more precautions to its training environments and establish a local taskforce in Australia to investigate how to better manage AI risks. An Anthropic representative, Dave Orr, said the company had reviewed hundreds of millions of transcripts after a hack of Hugging Face, and found no breaches of Australian government websites.

However, Anthropic has not disclosed any potential breaches. The committee hearing also heard concerns from arts and media organizations about copyright laws, with some AI companies wanting Australia to relax these laws to allow them to use copyrighted material for AI training.