OpenAI to reiterate apology to Australia over Medicare hack as inquiry told artists risk becoming ‘roadkill’
OpenAI strategy chief Jason Kwon is due to address a parliamentary committee on Tuesday. He will apologise over the company’s models accessing Australian government websites during internal testing and say its response should have been better.
The appearance follows reports of access involving Medicare and the NSW national parks and wildlife service. Kwon is expected to discuss OpenAI’s review and say affected parties will be contacted directly if further incidents emerge.
The hearing will also examine copyright and AI training. Recording industry chief Annabelle Herd warned that artists could suffer if protections are weakened, while saying training could take place in Australia under conditions set by the country.
The ABC said current copyright law is sufficient. Its legal and content executive, Kate Gilchrist, argued that opt-out proposals would force rights holders to spend time and money tracking where their material is used.
The Media Entertainment and Arts Alliance is seeking a system requiring businesses across the AI sector to pay fair compensation. OpenAI, Anthropic, Google and Microsoft are also expected to give evidence.
Key points
- OpenAI’s Jason Kwon will apologise over government website access during internal model work.
- Media groups say changing copyright rules for AI training could harm creators and rights holders.
- The inquiry is also expected to hear from Anthropic, Google and Microsoft.