European Commission introduces new channels for reporting AI Act violations
The European Commission’s AI Office has launched several tools to support enforcement of the EU AI Act, including a complaint mechanism for alleged violations, a dedicated channel for downstream providers of general-purpose AI models, and a whistleblower tool for people working with AI providers.
The European Commission’s AI Office has introduced new channels through which individuals, organisations, and people working in the AI sector can report potential violations of the EU AI Act.
The main mechanism is a Complaint Tool, which allows natural and legal persons to report alleged infringements by providers or deployers of AI systems that fall under the AI Office’s exclusive enforcement responsibilities. The mechanism is based on Article 85 of the AI Act.
The tool has a defined scope. Complaints can only concern matters for which the AI Office has enforcement responsibility under the AI Act. It is therefore not intended for reporting every possible problem involving an AI system or alleged violations of other EU or national laws.
A separate channel is available for downstream providers that work with general-purpose AI models. Under Article 89(2), these providers can raise concerns when they believe a general-purpose AI model provider has failed to meet certain obligations. These include requirements related to technical documentation, copyright policies, summaries of training data, incident reporting, and cybersecurity. For models classified as having systemic risks, the relevant obligations also include risk assessment and mitigation.
The AI Office has also established a Whistleblower Tool for people working with AI providers who want to report suspected violations. The tool is intended to provide a more secure channel for individuals who may have information about potential non-compliance within an organisation.
The new mechanisms are part of the EU’s developing enforcement framework for the AI Act. Rather than relying only on regulators to identify possible violations, the channels provide different groups with ways to bring potential non-compliance to the attention of the AI Office, depending on their role and the type of obligation involved.
The tools do not replace national enforcement mechanisms. The AI Act assigns different responsibilities to the European Commission, the AI Office, national competent authorities, and other bodies, meaning that the appropriate reporting channel depends on the type of AI system and the specific requirement that may have been breached.
