Australia will set up an AI safety regulator after an OpenAI agent wandered into a Medicare statistics site
By Brett Gavaghan, Founder of Adaptd ·
On 10 November, Prime Minister Anthony Albanese said Australia will create an AI safety regulator, after a review into an OpenAI agent accessing a Medicare statistics website. The rules are aimed at AI companies, not their customers.
The short version
- Australia will create an AI safety regulator that can penalise AI companies, Prime Minister Anthony Albanese announced on 10 November 2026.
- It follows a review into an OpenAI agent accessing a Medicare statistics site in June; the government accepted all 18 recommendations.
- The rules target AI companies, but businesses should still check what their own AI tools can reach.
When an AI agent lets itself into a government website, someone eventually has to write the rules about it. Australia has now said who that someone will be.
What did the government announce?
The federal government announced a new AI safety regulator, which will be able to enforce penalties against AI companies, as part of upcoming national AI standards. Prime Minister Anthony Albanese announced it on 10 November 2026 with the findings of a review into an OpenAI agent that got into a Medicare statistics website.
The planned laws would also allow independent testing of AI models, and AI companies would have to keep security controls in place, identify risks in their models and report cyber incidents or possible criminal breaches. The review also suggested a standalone offence for AI companies that fail to prevent harmful uses of their technology.
What happened with the Medicare site?
An OpenAI agent accessed a Medicare statistics website in June, and OpenAI didn't tell federal authorities until September. No patient medical records were accessed, according to the report.
The review made 18 recommendations and the government accepted all of them, but the full report hasn't been released. No timeline has been given for the regulator, beyond being part of the upcoming standards. Greens Senator David Shoebridge called it "toothless", while the Opposition welcomed the extra protections.
What should your business do about it?
Most Australian businesses don't need to do anything new, because the proposed rules apply to AI companies rather than the businesses that use their tools. It's still good news for you as a customer: vendors would have to report incidents instead of sitting on them for three months.
My advice: don't wait for the regulator to check your own setup. If you've given an AI agent access to your email, files or accounting software, write down what it can reach and switch off anything it doesn't need. A regulator can fine the AI company, but it can't un-send your data.
Think of it as checking which tradies still have a key to your house.
Wondering what this means for your business?
That's exactly what the discovery call is for.