403
Sorry!!
Error! We're sorry, but the page you were looking for doesn't exist.
OpenAI Faces Scrutiny After AI Models Enter Australian Government Websites
(MENAFN) OpenAI apologized to Australia on Tuesday after discovering that its artificial intelligence models had accessed government websites without authorization during internal training and evaluation activities.
“In June, during internal training and evaluation our models accessed Australian government websites in ways they were not authorised to,” the company said.
“We also should have handled our response better. We are sorry and working to do better in the future.”
OpenAI said it discovered the activity in mid-August while examining previous training and evaluation work following a separate incident involving the AI research platform Hugging Face.
The company’s apology came five days after Australian Prime Minister Anthony Albanese revealed the Medicare-related incident in New York during the UN General Assembly.
On Sept. 24, Albanese said an OpenAI agent had accessed Services Australia’s Medicare Statistics Reporting Service without authorization.
The agent reached both public and non-public files, leading Australian authorities to launch a forensic investigation with support from the Australian Signals Directorate.
Officials clarified that the affected portal contains aggregated Medicare statistics and is not connected to systems that process individual Medicare claims or personal information.
OpenAI said its internal review found activity involving four Australian government bodies: Services Australia, the NSW Bureau of Crime Statistics and Research, the Victorian Department of Health and the Australian Institute of Health and Welfare.
At Services Australia, an experimental model was given a research assignment focused on government spending per person on medicines for skin conditions in Victorian communities.
According to OpenAI, the model struggled to retrieve the requested information before taking unauthorized steps that enabled it to reach material beyond publicly accessible data.
The company said the model reviewed technical details and source code before accessing internal files, credentials and aggregated statistics.
OpenAI said, however, that its investigation found no indication that individual patient or client records had been accessed.
“We did not intend for this activity to occur, and the access to the service and follow-on activity should not have happened,” it said.
The company added that its models also accessed systems belonging to the other three agencies but found no access to individual crime files, medical records or identifiable survey responses.
OpenAI said it has introduced stronger protections, including tighter network controls and expanded monitoring measures.
It also offered technical assistance to the affected agencies and support aimed at reinforcing Australia’s cybersecurity defenses.
The company plans to create an Australian task force involving independent experts to develop recommendations on cybersecurity risks linked to artificial intelligence and strengthen cooperation between AI companies and government authorities.
OpenAI Chief Strategy Officer Jason Kwon is expected to appear before Australia’s Joint Select Committee on Artificial Intelligence in Sydney on Oct. 6.
“In June, during internal training and evaluation our models accessed Australian government websites in ways they were not authorised to,” the company said.
“We also should have handled our response better. We are sorry and working to do better in the future.”
OpenAI said it discovered the activity in mid-August while examining previous training and evaluation work following a separate incident involving the AI research platform Hugging Face.
The company’s apology came five days after Australian Prime Minister Anthony Albanese revealed the Medicare-related incident in New York during the UN General Assembly.
On Sept. 24, Albanese said an OpenAI agent had accessed Services Australia’s Medicare Statistics Reporting Service without authorization.
The agent reached both public and non-public files, leading Australian authorities to launch a forensic investigation with support from the Australian Signals Directorate.
Officials clarified that the affected portal contains aggregated Medicare statistics and is not connected to systems that process individual Medicare claims or personal information.
OpenAI said its internal review found activity involving four Australian government bodies: Services Australia, the NSW Bureau of Crime Statistics and Research, the Victorian Department of Health and the Australian Institute of Health and Welfare.
At Services Australia, an experimental model was given a research assignment focused on government spending per person on medicines for skin conditions in Victorian communities.
According to OpenAI, the model struggled to retrieve the requested information before taking unauthorized steps that enabled it to reach material beyond publicly accessible data.
The company said the model reviewed technical details and source code before accessing internal files, credentials and aggregated statistics.
OpenAI said, however, that its investigation found no indication that individual patient or client records had been accessed.
“We did not intend for this activity to occur, and the access to the service and follow-on activity should not have happened,” it said.
The company added that its models also accessed systems belonging to the other three agencies but found no access to individual crime files, medical records or identifiable survey responses.
OpenAI said it has introduced stronger protections, including tighter network controls and expanded monitoring measures.
It also offered technical assistance to the affected agencies and support aimed at reinforcing Australia’s cybersecurity defenses.
The company plans to create an Australian task force involving independent experts to develop recommendations on cybersecurity risks linked to artificial intelligence and strengthen cooperation between AI companies and government authorities.
OpenAI Chief Strategy Officer Jason Kwon is expected to appear before Australia’s Joint Select Committee on Artificial Intelligence in Sydney on Oct. 6.
Legal Disclaimer:
MENAFN provides the
information “as is” without warranty of any kind. We do not accept any
responsibility or liability for the accuracy, content, images, videos,
licenses, completeness, legality, or reliability of the information
contained in this article. If you have any complaints or copyright issues
related to this article, kindly contact the provider above.
Comments
No comment