New: Had a problem with an AI system? Australia now has a place to report it.
Report an AI incident

The AI Incident Exchange

Terms of submission

Last updated 12 June 2026

These terms apply when you submit a report to the AI Incident Exchange operated by Responsible AI Australia (RAIA). By submitting a report you accept them. They sit alongside our terms of service and privacy policy.

1. What this service is, and is not

The AI Incident Exchange records and analyses reports of incidents and near-misses involving AI systems, so that patterns of harm can be identified and published in de-identified, aggregate form. RAIA is not a regulator, ombudsman, complaint scheme or dispute resolution body. Submitting a report does not lodge a complaint with any organisation, does not start any legal process, and cannot result in compensation, correction of a record, or any other personal outcome. Nothing on this service is legal advice.

If you need an outcome, bodies with the power to help include Scamwatch (scams and fraud), the Australian Financial Complaints Authority (financial services disputes), the Office of the Australian Information Commissioner (privacy), and the eSafety Commissioner (online harms). If anyone is in immediate danger, call 000.

2. Your declarations

By submitting a report, you declare that:

  • the report is true to the best of your knowledge;
  • it describes something you experienced, witnessed, or were told about by the person affected, and your answer about your connection to the incident is accurate;
  • you have not included another person's personal information beyond what is necessary to describe the incident, and you have not included material you are not permitted to share (for example, material subject to confidentiality obligations or court orders);
  • you understand the report is information for analysis, not an allegation RAIA has verified or adopted.

3. What we do with reports

Reports are reviewed by a person, classified against Australia's AI Ethics Principles and common harm types, and de-identified for analysis. We publish findings only in aggregate form: categories, frequencies and trends. We do not publish individual reports, the text of any report, your identity, or the identity of any person or business named in a report. We may share de-identified, aggregate data with researchers and government bodies working on AI safety.

An individual report may be referred, with your consent, to an appropriate authority where it describes serious harm. We will only contact you if you gave us permission and provided an email address.

4. Privacy

Your report may contain personal information. We collect it to operate the AI Incident Exchange as described in these terms, handle it in accordance with our privacy policy, and do not use your contact details for marketing. You may ask us to correct or delete your report at any time using the contact details below.

5. Corrections and withdrawal

If you need to correct a report, withdraw it, or believe a report concerns you or your business and is wrong, contact us through the contact page. We maintain a documented correction and takedown procedure and will respond promptly.

6. Changes

We may update these terms from time to time. The date above shows the current version. Reports are governed by the terms in force when they were submitted.