OpenAI to apologize again to parliamentary inquiry over AI agent's access to Australian government websites

The Guardian reports in a preview that OpenAI Chief Strategy Officer Jason Kwon will apologize again at a hearing of Australia's Joint Parliamentary Committee on Artificial Intelligence over the company's models accessing government websites during internal training and evaluation. The report quotes his statements that the company should have handled the response better, and that if ongoing reviews found more incidents it would notify relevant parties more promptly and directly; previously the government was notified only three months after the incident occurred. The report also discusses the controversy over copyright licensing for AI training. The article expects OpenAI, Anthropic, Google and Microsoft to testify on Tuesday afternoon; this cannot confirm that the testimony or the renewed apology has actually occurred.