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

Our standing commitment

Independence and impartiality

A certification is only worth something if the body behind it is independent. Responsible AI Australia is a certification body, not a consultancy. This page sets out exactly how we keep our judgement arms-length from commercial interest, so the mark on the register means what it says.

Our commitments

We do not consult for the businesses we certify

Responsible AI Australia does not sell consulting, remediation or advisory services. The body that judges a business must not also be paid to help it pass. This is the single most important line we hold.

Decisions follow published criteria

Certification is granted on the evidence a business provides, assessed against our published tier criteria and the Australian AI Ethics Principles. The outcome does not depend on who the applicant is or what they pay beyond the standard certification fee.

We identify and manage conflicts of interest

Before any assessment, we check for relationships that could compromise impartiality. Where an assessor has a personal or commercial connection to an applicant, they recuse themselves and an alternative reviewer is assigned.

Our funding is transparent

Our revenue comes from certification and from the public register, not from helping businesses prepare for the very assessment we run. Removing the advisory line removes the incentive to pass a business that should not pass.

Certification can be challenged and revoked

Anyone can raise a concern about a certified business through our complaints process and the AI Incident Exchange. A certified business must respond within 28 days, and unresolved issues can lead to revocation. Certification is a standing claim, not a one-time sale.

Help to prepare is free and open to all

We publish free guidance, the Responsible AI Pledge and the Australian AI Ethics Principles so any business can prepare on its own terms. That support is the same for everyone and is never a paid path to a better result.

How we manage conflicts of interest

Because we no longer offer advisory services, the most serious conflict, advising a business and then certifying it, cannot arise through our own work. Any business that received advisory work from us in the past is subject to an independent review and a cooling-off period before it can be certified, so no historical relationship shortcuts an assessment.

For every assessment, the reviewer confirms they have no personal or commercial relationship with the applicant. Where one exists, the reviewer steps aside and the assessment is reassigned. Borderline certification decisions are reviewed by a second person before the mark is issued.

Aligned with international principles

Our approach is modelled on the principles that govern recognised certification bodies internationally, including ISO/IEC 17065 and ISO/IEC 17021:

  • Impartiality: assessment free from commercial and personal bias
  • Competence: decisions made by people qualified to make them
  • Responsibility: the applicant remains responsible for its own AI practices
  • Openness: criteria, tiers and this policy are public
  • Confidentiality: applicant information is handled in confidence
  • Responsiveness: complaints and appeals are taken seriously and acted on

Think we have fallen short?

Independence only matters if it can be challenged. If you believe a certification decision was not impartial, tell us.

Raise a concern

Responsible AI Australia (ABN 16 675 184 291). Last updated 14 June 2026.