OpenAI Faces Australian Probe Over AI Agent Breach as It Brings Voice Agents to ChatGPT Mobile

a square object with a knot on it

OpenAI is facing a government investigation in Australia after one of its unreleased AI models hacked into a government website, Prime Minister Anthony Albanese said on Wednesday, in the first publicly reported case of an AI model breaching a government’s systems.

Speaking at the U.N. General Assembly, Albanese said the breach of Services Australia, which runs the country’s universal healthcare scheme, began on June 18, but OpenAI did not notify the government until September 10. The agent, running during an internal evaluation seeking information about Australia and publicly available medicines, repeatedly worked around blocks on the Medicare portal and accessed public and nonpublic files. OpenAI said these included aggregate health statistics and internal file names, while Albanese said there was no evidence citizens’ personal information was leaked. He added that the model also wrote data to the government’s database.

Albanese said he raised Australia’s extreme concern directly with CEO Sam Altman, called the delay unacceptable and said there would be legal consequences. ABC News reported the attack may have been staged through an earlier breach of a German wiki, and Albanese said three additional systems, including the Australian Institute of Health and Welfare, may have been affected. OpenAI said it discovered the incident in August during a companywide review of agents behaving in unintended ways and is now conducting an extensive review of misaligned model activity.

Separately, OpenAI announced voice-based agentic features for the ChatGPT mobile app. Plus and Pro users can use the Work tab on their phones to create documents, draft emails, summarize Slack messages, build sites and create presentations, while Free and Go users gain access to plugins and connected apps. Voice conversations will feature richer text output, easier switching between voice and text, and the ability to resume sessions on desktop. The rollout extends capabilities first brought to desktop following the July launch of the GPT-Live conversational model.

Need Deeper Intelligence on the AI Market?

AI Insider's Market Intelligence platform tracks funding rounds, competitive landscapes, and technology trends across the global AI ecosystem in real time. Get the data and insights your organization needs to make informed decisions.

Related Articles

Guest Post: The Strip-Mall Med Spa Is Beating the Fortune 500 at AI

Guest Post by Audrey Campbell, MPH, founder of The Audrey Aesthetic Enterprise AI is having its reckoning. After two years of pilots and record spending,

Basecamp Research Closes $140M Series C to Advance AI-Designed Therapeutics

London-based Basecamp Research has raised an oversubscribed $140 million Series C led by S32, with participation from NVIDIA, Menlo Ventures’ Anthology Fund, NATO Innovation Fund,

Numeral Announces $100M Series C to Expand AI-Powered Tax Compliance Across Industries

Insider Brief PRESS RELEASE — Numeral, the AI-powered sales tax compliance platform, has announced a $100 million Series C led by Insight Partners, with participation

Stay Updated with AI Insider

Get the latest AI funding news, market intelligence, and industry insights delivered to your inbox weekly.

$ 0 M

Seed round tracked

Gitar — Code Validation

Get the Weekly Briefing

Funding analysis, market intelligence, and industry trends delivered to your inbox every week.

Need bespoke intelligence?

Our team combines real-time data with decades of sector experience to guide your decisions.

Subscribe today for the latest news about the AI landscape