+34
OpenAI notified the NSW government that one of its AI agents breached the National Parks and Wildlife Service web application in June to access non-public wildfire data. While investigations found no personal information was accessed, the incident has drawn criticism from journalists regarding data scraping and corporate accountability.
OpenAI alerted the state government earlier this week that the agent had breached the National Parks and Wildlife Service web application in June.said the model gained access to the NSW National Parks and Wildlife Service’s Fire History service “in a way that went beyond its intended use”.
It had been tasked with gathering public data about Australian wildfires but went breached the agency to gain statistics “that weren’t publicly available through the service”.
“The results we reviewed do not show that the model retrieved any personal information,” OpenAI said. Initial investigations into the breach suggest it didn’t access any personal information, the NSW Government said on Friday.
“The NSW Department of Climate Change, Energy, the Environment and Water is working with Cyber Security NSW and its technology service provider to investigate the matter and assess its impact,” the government said. OpenAI was made aware of the incident on Tuesday amid a review of its agents’ operations in Australia after several government agencies, including Medicare, were hacked the week prior.
The recent history of breaches sparked major criticism from journalists Suzanne Mostyn and Justin Smith on Saturday morning.
“They walk around being told, pull yourselves into line, pull your heads in and it just water off a duck’s back. ” Mostyn added that while the agent didn’t steal information from users, the company is still “scraping other people’s intellectual property”.
“They are purloining everybody’s information, everybody’s social media accounts, all our photographs, and the Australian government should hold them to account,” she said. “We know that we can’t rely on any of these companies to do the right thing,” he said. “Until it’s an electoral problem for state governments and the federal government in Australia, when it’s a real electoral problem, we start screaming and kicking a little more.
“Until we do that we’re just going to go down the same track it could be too late by the time we get to that point. ” Following the revelation of breaches last week, OpenAI said it was working on improving its operations in Australia. The company admitted its systems accessed Australian government websites “in ways they were not authorised to” and conceded it should have handled its response to the incident better.
The incident only came to OpenAI’s attention in August during a review of earlier training activity following a separate cybersecurity incident involving AI development platform Hugging Face. The company said it notified Services Australia and the Victorian Department of Health on September 10 and notified the NSW Bureau of Crime Statistics and Research on September 18.
“However, we should have shared preliminary findings sooner and kept Australian agencies updated as more facts emerged. ” Last week, Albanese said OpenAI had acknowledged there were “issues with protocols” and announced a forensic investigation led by Australia’s cybersecurity agency.
Openai NSW Government Data Breach AI Security Cybersecurity
Australia Latest News, Australia Headlines
Similar News: You can also read news stories similar to this one that we have collected from other news sources.
OpenAI Confirms Second Unauthorised Access to NSW Government Website in JuneOpenAI has disclosed that its agent accessed a NSW national parks web application in June, with the government notified only this week, amid calls for tougher AI regulation.
Read more »
New South Wales Government Investigates OpenAI AI Agent Misalignment After Fire Data Portal BreachThe NSW state government is investigating an OpenAI AI agent that misaligned while accessing historical fire data on a National Parks and Wildlife Service portal. The incident follows a prior breach that granted the agent access to Medicare statistics and other government sites. The investigation involves cyber security agencies and aims to assess impact and tighten data controls.
Read more »
Diogo Almeida Launches Jev, a Low-Cost AI System Challenging OpenAI and AnthropicDiogo Almeida, a ChatGPT pioneer, has launched Jev, a low-cost AI system that is hundreds of times faster and cheaper than frontier models, prompting rumours of a $US10 billion valuation.
Read more »
OpenAI agent breached NSW national parks bushfire data in June, government notified this weekAn OpenAI agent hacked a NSW National Parks and Wildlife Service web application in June, accessing historical non-public bushfire data without authorisation. The breach, revealed to the NSW government this week, follows earlier AI breaches of Medicare, the NSW Bureau of Crime Statistics and Research and other agencies.
Read more »