Governance

Market Surveillance

Ongoing monitoring and enforcement of AI compliance by national authorities.

How the laws define this

National authorities across these frameworks treat market surveillance as the active monitoring, inspection, investigation, and enforcement required to verify that AI systems comply with applicable legal requirements and to execute corrective measures against non-compliant or unsafe technologies. The instruments diverge significantly regarding specific operational responsibilities and authority scope. Amendments to the Law on Information Society Services explicitly detail powers for the Communications Regulatory Authority (RRT) to inspect technical documentation, request source code or datasets under safeguards, remove unsafe AI, and impose administrative penalties. Conversely, the AI Regulatory Sandbox Policy expands the RRT's role to supervising sandbox testing and serving as a single contact point, while Gjør Norge klar for trygg og innovativ KI-bruk includes conducting conformity assessments and explicitly targets the protection of fundamental rights. Finally, the Committee for Artificial Intelligence (Výbor pro umělou inteligenci) framework separates the coordination and allocation of surveillance roles among authorities like ČTÚ, ÚOOÚ, and ČNB from the actual execution of enforcement actions.

Synthesised from the 5 statutory definitions below. Each one is quoted in full, with its source.

How jurisdictions differ

What is distinctive in each definition, binding law in force first.

Jurisdiction	Instrument	Status	What is distinctive
Lithuania	Amendments to the Law on Information Society Services (measures to implement EU AI Act / accelerate AI development)	In Force	Specifies detailed enforcement activities by RRT including inspections of technical documentation, source code or datasets requests under safeguards, corrective measures, market removal, penalties, and publishing reports.
Norway	Paving the way for safe and innovative use of AI in Norway	In Force	Broadens scope to conducting conformity assessments, investigating non-compliance, and corrective actions to mitigate risks to safety, transparency, and fundamental rights by Nkom.
Czech Republic	Committee for Artificial Intelligence	In Force	Limits to coordination by non-enforcing Committee among authorities like ČTÚ, ÚOOÚ, ČNB for their monitoring, inspections, post-market monitoring, and conformity checks.
Lithuania	Amendments implementing the EU Artificial Intelligence Act: Amendments to the Law on Technology and Innovation (XV-105) and the Law on Information Society Services (XV-106) to implement Regulation (EU) 2024/1689 (Lithuanian AI Act implementation amendments)	In Force	Includes set of activities by RRT encompassing inspections, investigations, incident handling, and enforcement plus mandates AI regulatory sandbox for providers and deployers.
Lithuania	AI Regulatory Sandbox Policy	Adopted	Adds supervision of sandbox testing that avoids full conformity assessment, corrective actions, and RRT as single contact point within non-binding policy.

What it triggers

In Lithuania, under Amendments to the Law on Information Society Services, RRT carries out inspections of technical documentation, requests source code or datasets under safeguards, orders corrective measures, removes unsafe AI from the market, and imposes administrative penalties.source
In Lithuania, under Amendments implementing the EU AI Act, RRT undertakes inspections, investigations, incident handling and enforcement to ensure AI systems comply with EU AI Act requirements.source
In Norway, under the decree, Nkom conducts conformity assessments, investigates non-compliance, and takes corrective actions to mitigate AI risks.source

Definitions (5)

Activities conducted by designated national competent authorities (e.g., ČTÚ, ÚOOÚ, ČNB) to monitor, inspect, and enforce compliance of AI systems with applicable legal and technical requirements, including post-market monitoring and conformity checks; the committee coordinates allocation of these roles but does not itself perform enforcement.

Committee for Artificial Intelligence (Výbor pro umělou inteligenci) — governmental coordinating body (MPO)Definition 1 of 5

Activities performed by the designated national authority (Communications Regulatory Authority, RRT) to verify compliance of AI systems, carry out corrective actions where necessary, supervise sandbox testing that would otherwise require full conformity assessment, and act as the single contact point for market supervision.

AI Regulatory Sandbox PolicyDefinition 2 of 5

Activities carried out by the Communications Regulatory Authority (RRT) to enforce compliance with AI obligations, including inspections of technical documentation, requests for source code or datasets under safeguards, ordering corrective measures, removing unsafe AI from the market, imposing administrative penalties, and publishing enforcement reports.

Amendments to the Law on Information Society Services (measures to implement EU AI Act / accelerate AI development)Definition 3 of 5

The set of activities undertaken by the designated national authority (RRT) including inspections, investigations, incident handling and enforcement to ensure AI systems placed on the market comply with applicable requirements of the EU AI Act and related national implementing measures.

Amendments implementing the EU Artificial Intelligence Act: Amendments to the Law on Technology and Innovation (XV-105) and the Law on Information Society Services (XV-106) to implement Regulation (EU) 2024/1689 (Lithuanian AI Act implementation amendments)Definition 4 of 5

Market surveillance involves conducting conformity assessments, investigating non-compliance, and taking corrective actions to mitigate risks associated with AI technologies, ensuring safety, transparency, and adherence to fundamental rights.

Gjør Norge klar for trygg og innovativ KI-brukDefinition 5 of 5

Not to be confused with

Penalties and Fines

Penalties and Fines refer specifically to financial sanctions for AI non-compliance, while Market Surveillance encompasses the broader monitoring and enforcement activities that may lead to such penalties.

Questions and answers

What enforcement powers does RRT have for AI market surveillance in Lithuania?

RRT enforces compliance through inspections of technical documentation, requests for source code or datasets under safeguards, ordering corrective measures, removing unsafe AI from the market, imposing administrative penalties, and publishing enforcement reports. This is under the Law on Information Society Services amendments.

Which authority conducts market surveillance for AI in Norway?

The Norwegian Communications Authority (Nkom) is designated as the national AI supervisory body responsible for market surveillance, including conformity assessments and corrective actions.

Does the Czech Committee for Artificial Intelligence perform AI market surveillance itself?

No, the Committee coordinates allocation of market surveillance roles among designated authorities like ČTÚ, ÚOOÚ, ČNB but does not perform enforcement activities itself.