EU opens consultation on AI incident reporting rules

Anyone interested, companies, civil society groups, or individuals, can submit feedback through the consultation portal until 7 November.

EU opens consultation on AI incident reporting rules

The European Commission has launched a public consultation on new rules for reporting serious incidents involving artificial intelligence systems. The consultation opened on 26 September and will run until 7 November 2025, giving stakeholders an opportunity to provide feedback on draft guidance and a reporting template.

Why it matters: The rules will affect not only AI providers but also the wider public. Civil society involvement is crucial to ensure that reporting obligations address broader concerns such as transparency, accountability, and human rights. By contributing, civil society can help shape rules that protect individuals and communities, not just companies and regulators.

The measures stem from the EU’s Artificial Intelligence Act, which introduces strict obligations for providers of high-risk AI systems. Under Article 73 of the Act, companies must notify national authorities when serious incidents occur, such as cases where AI systems cause harm, malfunction, or present significant risks. The aim is to detect problems early, ensure accountability, and build public trust in the technology. Although these rules will only take effect from August 2026, the draft documents are already available to help companies prepare.

The draft guidance explains key definitions, gives practical examples, and clarifies how the new requirements relate to existing EU and national reporting regimes. It also highlights efforts to align with international initiatives, including the OECD’s AI Incidents Monitor and Common Reporting Framework.

Civil society groups representing consumers, human rights, and digital rights play a key role in this process. Their input can help ensure that reporting rules are workable in practice while also safeguarding people from harm. By raising issues of accessibility and fairness, they can influence how AI accountability is designed and monitored in Europe.

Anyone interested, companies, civil society groups, or individuals, can submit feedback through the consultation portal until 7 November.

Go to Top