Glossary · EU AI Act: high-risk AI
Post-market monitoring (AI Act)
Also known as: Post-market monitoring system (AI Act), Post-market monitoring plan
German: Beobachtung nach dem Inverkehrbringen (KI-Verordnung)
Under the EU AI Act, post-market monitoring is the system a provider of a high-risk AI system establishes and documents to actively and systematically collect, document and analyze data on the system’s performance throughout its lifetime — including data provided by deployers — so that it can evaluate continuous compliance with the high-risk requirements.
- AI regulation
- EU
In one sentence
Post-market monitoring under Art. 72 AI Act: the provider’s documented system for collecting and analyzing data on a high-risk AI system in use.
Example
The provider of an AI-based visual inspection system collects false-reject rates and operator overrides from deployers every month and reviews them against the accuracy stated in the instructions for use.
Explained in context
This term is a station of a mini course
- The medical-device evidence chain: from intended purpose to post-market evidenceStation: Real-world performanceStart the course
Context cards connect this term with others to answer one question. Also in British English and German.
- One technical file, two sets of requirements: what the AI Act adds to an MDR or IVDR fileDoes a medical device with AI need a second technical file for the EU AI Act?UKDeutsch
- The medical-device evidence chain: from intended purpose to post-market evidenceHow do the parts of a medical device’s technical documentation connect — and what has to be in place before an AI function may answer from it?UKDeutsch
- Reporting duties after an AI agent incident: AI Act, CRA, NIS 2 and GDPRWhen an AI agent has caused an incident, which EU reporting duties can apply — who reports, to whom, and by when?UKDeutsch
How it applies
- Provider's system: Proportionate to the nature of the AI technology and the risks; based on a post-market monitoring plan that is part of the technical documentation (Annex IV). Where relevant it includes the interaction with other AI systems.
- Data sources: Data that deployers provide or that is collected otherwise — performance, deviations, overrides, logs where the provider has access (record-keeping and logs).
- Feeds incident reporting: Post-market monitoring is where many serious incidents are first recognized; the two are separate duties.
- When it applies: With the high-risk obligations — December 2, 2027 for Annex III systems and August 2, 2028 for Annex I Section A products, as set by the Digital Omnibus on AI.
- AI agents: For an agent within a high-risk system, monitoring covers its actions, not only its answers: tool calls, refusals, approvals and interventions by people.
Post-market monitoring vs. post-deployment monitoring
Post-deployment monitoring is the general engineering practice of watching a model in operation. Post-market monitoring under Article 72 is the provider's legal duty for high-risk AI systems, with a documented plan. The practice can serve the duty; it doesn't establish compliance by itself.