Already devastated by record-breaking floods that claimed 13 lives and destroyed thousands of homes, Northern Rivers residents now face a fresh betrayal of trust. In March, a former NSW government contractor uploaded the personal details of 3,000 flood victims to ChatGPT—the same artificial intelligence platform that millions use for casual conversations and homework help.
The breach occurred between March 12-15, 2025, when personal information from the Northern Rivers Resilient Homes Program was uploaded to the AI platform by a former contractor.
Email addresses, phone numbers, and 'other personal and health information' of people seeking help to rebuild their lives are now in the hands of a commercial AI company with no guarantee they can be retrieved.
'We understand this news is concerning and we are deeply sorry for the distress it may cause for those who have engaged with the program.'
The unsettling details of what was shared
The uploaded data included a Microsoft Excel spreadsheet containing 10 columns and more than 12,000 rows of information—a staggering amount of personal detail that authorities are still working through.
While the NSW Reconstruction Authority insists there's 'no evidence' the information has been accessed by third parties, they admit this possibility 'cannot be ruled out.'
This isn't the first time Australian government workers have misused ChatGPT with sensitive data. Victoria's child protection agency was recently found to have breached privacy laws when a worker entered 'significant amounts of personal and delicate information' about an at-risk child into the platform.
What is ChatGPT?
ChatGPT is an artificial intelligence chatbot that can answer questions, write content, and help with various tasks.
When you type something into ChatGPT, that information becomes part of its database and could potentially be used to train future versions of the system.
This is why uploading personal or sensitive information is so problematic—once it's in, it's extremely difficult to remove.
A pattern of government AI misuse emerges
Recent research shows 77 per cent of companies are using or considering AI tools like ChatGPT in their business operations, but Australian government departments seem to be struggling with proper protocols.
Victoria's Department of Families, Fairness and Housing has now been ordered to block ChatGPT use by child protection workers following their privacy breaches.
The fundamental problem is that ChatGPT 'wasn't exactly built for enterprise use' and 'doesn't follow your security policies, respect your compliance boundaries, or ask permission before processing sensitive data.' Yet government employees continue to use it as if it were a secure, internal system.
The devastating floods that started this ordeal
To understand why this breach is particularly cruel, it's important to remember what these victims have already endured. In late February and early March 2022, the Northern Rivers region experienced record-breaking rainfall, with some areas receiving well over a metre of rain.
The Richmond, Hawkesbury-Nepean, Wilsons, Tweed, Brunswick, and Clarence rivers all flooded, inundating properties and whole townships.
More than 5,000 homes were damaged in NSW following the flooding, with some residents stepping 'out of their beds and into floodwater' due to communication outages that prevented proper warnings. Thirteen people lost their lives, and entire communities were torn apart.
The Northern Rivers Resilient Homes Program was meant to be a lifeline for these traumatised communities, offering financial support for home buybacks, raising, or retrofitting. Instead, those who trusted the system with their most vulnerable information have been let down again.
Did you know?
Did you know?
According to recent Australian data privacy research, the health services sector reported the highest number of data breaches at 104 incidents, but government departments are increasingly featuring in breach reports due to AI tool misuse.
What the government is doing to fix this mess
Since discovering the full scope of the breach, the NSW Reconstruction Authority says it has engaged forensic analysts and is working with Cyber Security NSW to investigate.
The breach has been reported to the NSW Privacy Commissioner in line with privacy laws, and the authority has 'reviewed and strengthened internal systems and processes and issued clear guidance to staff on the use of unauthorised AI platforms.'
Safeguards are now in place to prevent future uploads of personal information to ChatGPT and other AI platforms. The department promises to contact affected individuals this week and is monitoring the internet and dark web to see if any information appears online.
ID Support NSW, the state agency that helps people affected by data breaches, will provide assistance with compromised documents and identity restoration. The NSW Reconstruction Authority has also committed to compensating reasonable out-of-pocket expenses if identity documents need replacing.
What you should do if you're affected
If you applied for assistance through the Northern Rivers Resilient Homes Program, here's what experts recommend:
Steps to protect yourself
- Wait for official contact from the NSW Reconstruction Authority this week
- Don't click on links or provide information to anyone claiming to be from the program unless you're certain it's legitimate
- Monitor your bank statements and credit reports for unusual activity
- Consider placing a fraud alert on your credit file through credit reporting agencies
- Keep records of any additional costs incurred due to this breach for potential compensation
- Contact ID Support NSW on 1800 001 040 if you need immediate help with identity concerns
Remember that ChatGPT cannot delete individual prompts from its system, meaning the platform would need to delete all data to remove your specific information. This highlights why prevention is so much better than trying to fix these breaches after they happen.
The bigger picture for government AI use
While tools like ChatGPT and DALL-E are examples of 'general-purpose AI systems' that can be adapted for various uses, Australian privacy authorities are increasingly concerned about their misuse in government settings.
Security experts argue that winning organisations in 2025 won't be those 'playing whack-a-mole with AI tools' but rather 'the ones who set up invisible protections that let employees move fast without accidentally blowing holes in their security posture.'
The NSW breach serves as a wake-up call for all levels of Australian government. While the federal government introduced privacy reform legislation in 2024, implementing proper AI governance clearly requires more than just new laws—it needs cultural change and proper training.
What This Means For You
For the flood victims affected by this breach, the timing couldn't be worse. Already struggling to rebuild their lives after one of Australia's most devastating natural disasters, they now face the uncertainty of not knowing where their personal information might end up or how it might be used.
The NSW government's promise to 'restore trust' in the Resilient Homes Program will require more than apologies and compensation. It will need genuine accountability and assurance that the very people they're meant to help won't be betrayed again by careless handling of their most personal information.
Have you been affected by this data breach, or do you have concerns about government use of AI tools? We'd love to hear your thoughts and experiences in the comments below.
Original Article
https://www.news.com.au/technology/...d0777ebea5400090be402b8314dede?from=rss-basic
Resilient Homes Program data breach | NSW Government
Cited text: Between 12 and 15 March 2025, personal information held for the Resilient Homes Program (RHP) was uploaded to the AI platform ChatGPT by a former RA c...
Excerpt: The breach occurred between March 12-15, 2025, when personal information from the Northern Rivers Resilient Homes Program was uploaded to the AI platform by a former contractor.
https://www.nsw.gov.au/departments-...authority/resilient-homes-program-data-breach
OVIC finds department responsible for breaches of privacy through use of ChatGPT—Office of the Victorian Information Commissioner
Cited text: The content generated by ChatGPT and then used by the Child Protection worker when drafting the PA report contained inaccurate personal information—...
Excerpt: This isn't the first time Australian government workers have misused ChatGPT with sensitive data.
https://ovic.vic.gov.au/mediareleas...r-breaches-of-privacy-through-use-of-chatgpt/
Data Privacy Facts & Statistics Australia (2025) - Red Search
Cited text: According to recent research, 77 per cent of companies are using or considering AI (such as ChatGPT) in their business operations, while 83 per cent said AI is a top ...
Excerpt: Recent research shows 77 per cent of companies are using or considering AI tools like ChatGPT in their business operations
https://www.redsearch.com.au/resources/data-privacy-statistics-australia/
OVIC finds department responsible for breaches of privacy through use of ChatGPT—Office of the Victorian Information Commissioner
Cited text: The notice outlines six specific actions, including a requirement that DFFH blocks the use of ChatGPT and other similar tools by child protection work...
Excerpt: Victoria's Department of Families, Fairness and Housing has now been ordered to block ChatGPT use by child protection workers following their privacy breaches.
https://ovic.vic.gov.au/mediareleas...r-breaches-of-privacy-through-use-of-chatgpt/
ChatGPT Security Risks in 2025: A Guide to Risks Your Team Might Be Missing—Concentric AI
Cited text: ChatGPT wasn’t exactly built for enterprise use. It doesn’t follow your security policies, respect your compliance boundaries, or ask permission befor...
Excerpt: The fundamental problem is that ChatGPT 'wasn't exactly built for enterprise use' and 'doesn't follow your security policies, respect your compliance boundaries, or ask permission before processing sensitive data.'
https://concentric.ai/chatgpt-security-risks-in-2025-a-guide-to-risks-your-team-might-be-missing/
The 2022 NSW Floods
Cited text: In late February and early March 2022, persistent heavy rain in north-east NSW broke all records, with some areas recording well over 1 metre of rain....
Excerpt: In late February and early March 2022, the Northern Rivers region experienced record-breaking rainfall, with some areas receiving well over a metre of rain.
https://mhnsw.au/stories/general/2022-nsw-floods/
2022 Australian Flooding—Center for Disaster Philanthropy
Cited text: More than 20,000 homes and businesses flooded in Queensland, and more than 5,000 homes were damaged in NSW following heavy rains and flooding in Febru...
Excerpt: More than 5,000 homes were damaged in NSW following the flooding
https://disasterphilanthropy.org/disasters/2022-australian-flooding/
NSW Far North Coast and Northern Rivers flood impact research, March 2022
Cited text: Warnings in Mullumbimby were inhibited due to communication outages—meaning people stepped out of their beds and into floodwater.
Excerpt: some residents stepping 'out of their beds and into floodwater' due to communication outages that prevented proper warnings.
https://www.preventionweb.net/news/...thern-rivers-flood-impact-research-march-2022
Resilient Homes Program data breach | NSW Government
Cited text: Yes, in line with the Privacy and Personal Information Protection Act 1998, the breach was reported to the NSW Privacy Commissioner.The NSW Privacy Co...
Excerpt: The breach has been reported to the NSW Privacy Commissioner in line with privacy laws, and the authority has 'reviewed and strengthened internal systems and processes and issued clear guidance to staff on the use of unauthorised AI…
https://www.nsw.gov.au/departments-...authority/resilient-homes-program-data-breach
Resilient Homes Program data breach | NSW Government
Cited text: We have reviewed and strengthened internal systems and processes and issued clear guidance to staff on the use of non-sanctioned AI platforms. Safegua...
Excerpt: Safeguards are now in place to prevent future uploads of personal information to ChatGPT and other AI platforms.
https://www.nsw.gov.au/departments-...authority/resilient-homes-program-data-breach
Managing ChatGPT and Privacy in Australian Organisations | P108
Cited text: The platform’s FAQs state that it cannot delete individual prompts. This would mean Australian organisations would need to request all their data to b...
Excerpt: Remember that ChatGPT cannot delete individual prompts from its system, meaning the platform would need to delete all data to remove your specific information.
https://privacy108.com.au/insights/chatgpt-and-privacy/
Guidance on privacy and the use of commercially available AI products | OAIC
Cited text: [9] ChatGPT and DALL-E are examples of general-purpose AI systems (as well as being generative AI, see further below).
Excerpt: While tools like ChatGPT and DALL-E are examples of 'general-purpose AI systems' that can be adapted for various uses
https://www.oaic.gov.au/privacy/pri...the-use-of-commercially-available-ai-products
ChatGPT Security Risks in 2025: A Guide to Risks Your Team Might Be Missing—Concentric AI
Cited text: Security teams that win in 2025 aren’t the ones playing whack-a-mole with AI tools and avoiding AI governance; they’re the ones who set up invisible p...
Excerpt: Security experts argue that winning organisations in 2025 won't be those 'playing whack-a-mole with AI tools' but rather 'the ones who set up invisible protections that let employees move fast without accidentally blowing holes in their…
https://concentric.ai/chatgpt-security-risks-in-2025-a-guide-to-risks-your-team-might-be-missing/
Data Privacy Facts & Statistics Australia (2025) - Red Search
Cited text: The Australian government introduced the Privacy and Other Legislation Amendment Bill 2024 in February 2023, implementing 23 of the 25 legislative pro...
Excerpt: While the federal government introduced privacy reform legislation in 2024, implementing proper AI governance clearly requires more than just new laws—it needs cultural change and proper training.
https://www.redsearch.com.au/resources/data-privacy-statistics-australia/
Seniors Discount Club
Latest Articles
-
-
-
Lines blur between human care and artificial understanding: 'Dangerous'
- Started by Gian T
- Replies: 0
-
-
Fresh privacy breach by government contractors raises hard questions over AI use
- Started by Gian T
- Replies: 0
