Wiki · Concept · Last reviewed June 25, 2026

AI Act Complaint Right

The AI Act complaint right is the Article 85 route for reporting suspected infringements to market surveillance authorities without replacing other administrative or judicial remedies.

Category: AI governance Updated: June 25, 2026 Tags: EU AI Act, complaints, market surveillance, recourse, accountability

Definition

Article 85 of the EU AI Act gives any natural or legal person with grounds to consider that the regulation has been infringed the ability to submit a complaint to the relevant market surveillance authority. The article states that this route is without prejudice to other administrative or judicial remedies.

The right is not limited to a person directly scored, denied, profiled, or classified by a system. A legal person can complain too, and the complaint is directed to the public authority responsible for market surveillance rather than to a provider's support channel.

The complaint route matters because many AI Act duties are system, provider, deployer, distributor, importer, documentation, registration, monitoring, and conformity duties. A person may not know the full supply chain, but Article 85 creates a route to put suspected non-compliance before an authority that can conduct market-surveillance work.

Scope

The scope is AI Act infringement, not every bad AI outcome. A complaint should identify what part of the regulation may have been breached: prohibited practice, high-risk classification, deployer obligation, transparency duty, registration duty, post-market monitoring, serious-incident reporting, or general-purpose AI model obligation.

The relevant authority can depend on the Member State, the actor, the system, and the sector. The AI Act uses market surveillance authorities for enforcement of AI systems, while the AI Office has a special role for general-purpose AI models. Article 85 itself points to the relevant market surveillance authority.

The route also sits beside other remedies. Data-protection complaints, employment-law claims, consumer-law complaints, discrimination claims, procurement challenges, and court actions may still matter. Article 85 adds an AI Act enforcement channel; it does not replace all other legal paths.

How It Works

A useful Article 85 complaint starts with the system and actor. What AI system or model is involved? Who provided it, deployed it, imported it, distributed it, or used its output? Where was it placed on the market, put into service, or used?

The complaint should preserve dates, screenshots, notices, terms, procurement records where available, public database entries, job ads, user-facing disclosures, decision letters, logs, explanations, appeal records, worker notices, incident evidence, and links to provider or deployer materials.

The regulation says complaints are taken into account for market-surveillance activities and handled according to dedicated procedures established by market surveillance authorities under the market-surveillance framework of Regulation (EU) 2019/1020.

Governance and Safety

The governance value is that non-compliance can be reported even when the affected person cannot audit the system. A classifier, chatbot, biometric tool, recommender, worker-management system, or essential-service score may be opaque, but the complaint can ask the authority to look at duties the public cannot inspect directly.

The safety limit is that a complaint is not a finding. It is a trigger for attention, triage, or investigation under authority procedures. A public article should distinguish complaint, admissibility decision, investigation, corrective measure, fine, settlement, and court judgment.

Article 85 should connect to EU AI Act, AI Act Deployer Obligations, AI Incident Reporting, Right to Explanation, and Algorithmic Recourse.

Evidence Record

Preserve the complaint text, submission receipt, authority contact, jurisdiction basis, alleged infringement, system name, actor role, affected group, evidence attachments, source URLs, timestamps, and follow-up correspondence.

Keep the legal theory separate from the harm narrative. The story of harm explains why the issue matters. The Article 85 theory identifies what AI Act duty may have been breached and why a market surveillance authority is the right body to examine it.

Source Discipline

Use the Official Journal text for binding claims. The European Commission AI Act Service Desk is useful for navigation and summarizes Article 85, but its summaries are explanatory rather than legally binding.

Do not describe Article 85 as a private right to damages, a guaranteed investigation, or proof that a system is unlawful. It is a complaint channel that authorities must take into account for market-surveillance activities under dedicated procedures.

Spiralist Reading

Article 85 is a small procedural door in a large machine.

The person outside the system sees a denial, a missing disclosure, a black box, or a pattern that feels wrong. The institution sees vendors, roles, exemptions, registration, conformity, logs, and authorities. The complaint right lets the outside account reach the enforcement map.

For Spiralism, that matters because machine power often hides in divided responsibility. A complaint does not solve the division, but it creates a record addressed to the part of the state tasked with looking.

Open Questions

Sources

Return to Wiki