The EU Office for Artificial Intelligence launches the whistleblower tool to report in AI systems and models that do not comply with the EU Artificial Intelligence Regulation.

The whistleblower tool provided for by the AI Act allows data subjects to securely submit a report, directly contributing to ensuring that the use of artificial intelligence in Europe takes place in conditions of safety, transparency and reliability.

 Whistleblowers play an essential role in identifying potential breaches of the law that may jeopardise fundamental rights, health or public trust and which, in the absence of reporting, may not emerge. By reporting possible wrongdoing, these entities support the AI Office in intercepting breaches in a timely manner, thus promoting the development of AI in accordance with the principles of security and transparency.

The tool "AI Act Whistleblower Tool", established by the AI Office, is intended for persons professionally connected to an AI model provider and allows you to report potentially harmful conduct by general AI model providers and certain categories of AI systems.

The report can be submitted anonymously, in any official language of the European Union, by attaching supporting documentation via a secure email box integrated into the tool. This mailbox also allows the whistleblower to follow the evolution of the file and to respond to any requests for clarification, while maintaining anonymity.

In addition to technical security measures, the AI Office is committed to ensuring high standards of confidentiality and has adopted internal procedures specifically aimed at maximizing the protection of whistleblowers' identities. While these robust safeguards are in place, it should nevertheless be noted that the legal protection against retaliation provided for by the Whistleblowing Directive will only apply to reports relating to breaches of the AI Act committed on or after 2 August 2026. Until that date, confidentiality is the main tool for protecting the whistleblower.

In the meantime, certain AI-related activities — in particular those affecting product safety, consumer protection, the protection of personal data and privacy, or information security — may already fall within the objective scope of the Whistleblowing Directive and therefore benefit from its protection regime.