Emerging models of national competent authorities under the EU AI Act

Probably not a substantial claim: may not pass S3 (load-bearing).

This approach suggests a pragmatic strategy that capitalizes on established regulatory expertise and existing strong institutional frameworks (Winfield, 2018).

“This approach suggests a pragmatic strategy that capitalizes on established regulatory expertise and existing strong institutional frameworks (Winfield, 2018).”

• cites Winfield, 2018 sourcePartly supported

  In “Building the roadmap” (page not provided), the paper proposes regulating some robots through a body similar to the Civil Aviation Authority, supporting the idea of drawing on established regulatory models. The “Concluding discussion” (page not provided) calls for strong institutional frameworks, but does not establish that such frameworks already exist and can be used by the proposed approach. (LLM, conf=0.82)

  Experimental — accuracy not yet measured

May combine two claims.

While most EU countries have yet to designate their competent authorities, the emerging patterns from early adopters provide valuable insights into different implementation strategies.

“While most EU countries have yet to designate their competent authorities, the emerging patterns from early adopters provides valuable insights into different implementation strategies.”

May combine two claims. The quote only partly states this claim.

Cyprus divided responsibilities between two existing regulatory bodies, with the Commissioner for Communications serving as notifying authority, market surveillance authority, and single point of contact.

“The Commissioner for Communications serves multiple key roles: notifying authority, market surveillance authority, and the single point of contact”

• AssumptionRests on an assumption: Official documents and government websites accurately reflect Member States' actual institutional arrangements and implementation models.

May combine two claims.

Lithuania designated the Innovation Agency as notifying authority and the Communications Regulatory Authority (RRT) as market supervisor and single point of contact.

“The Innovation Agency has been designated as the national notifying authority, while the Communications Regulatory Authority will act as the market supervisor and single point of contact.”

• AssumptionRests on an assumption: Official documents and government websites accurately reflect Member States' actual institutional arrangements and implementation models.

May combine two claims.

Regulatory scope for AI oversight under the AI Act varies between horizontal and vertical frameworks, with some jurisdictions pioneering hybrid solutions.

“regulatory scope demonstrates a critical divide between horizontal and vertical oversight frameworks”

• AssumptionRests on an assumption: Official documents and government websites accurately reflect Member States' actual institutional arrangements and implementation models.

May combine two claims.

Italy proposed the National Cybersecurity Agency (ACN) as market surveillance authority and the Agency for Digital Italy (AgID) as notifying authority.

“proposed to appoint the National Cybersecurity Agency (ACN) as market surveillance authority with monitoring, inspection and enforcement powers, and the Agency for Digital Italy (AgID) as notifying authority.”

• AssumptionRests on an assumption: Official documents and government websites accurately reflect Member States' actual institutional arrangements and implementation models.

Probably not a substantial claim: may not pass S3 (load-bearing).

It is proposed to designate the Department of Digital Affairs of the Chancellery of the Prime Minister of Poland as the authority responsible for notification of conformity assessment bodies and as the notifying authority.

“It is also proposed to designate the Department of Digital Affairs of the Chancellery of the Prime Minister of Poland, the institution responsible for computerization, as the authority responsible for notification of conformity assessment bodies and as the notifying authority.”

• cites Government bill n.UC71 of 2024Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

Luxembourg designated the Commission for Data Protection (CGPD) as the horizontal market surveillance authority by default and single point of contact.

“the Commission for Data Protection (Commissariat du gouvernement à la protection des données auprès de l'Etat, CGPD) is responsible for coordination with market surveillance authorities and designated as the single point of contact.”

• AssumptionRests on an assumption: Official documents and government websites accurately reflect Member States' actual institutional arrangements and implementation models.

May combine two claims.

Hungary and Poland have opted to establish new AI-specific bodies.

“In contrast, Hungary and Poland have opted to establish new AI-specific bodies.”

• Also stated in: § 6. Conclusion · plain

  “Only Hungary and Poland opted to create new dedicated AI oversight bodies, highlighting different philosophies about building regulatory capacity.”

May combine two claims.

The fragmented approach emerges as the predominant choice among early implementers, with countries like Finland, Lithuania, Luxembourg, and Malta adopting various mixed models.

“The fragmented approach emerges as the predominant choice among early implementers, with countries like Finland, Lithuania, Luxembourg, and Malta adopting various mixed models.”

• Also stated in: § 6. Conclusion · plain

  “The fragmented approach emerges as predominant, with countries like Finland, Lithuania, Luxembourg, and Malta distributing responsibilities across multiple bodies.”

May combine two claims.

The results reveal a clear dichotomy in institutional approaches to AI oversight, with most Member States opting to leverage existing regulatory bodies while only a few create new AI-specific institutions.

“The results also reveal a clear dichotomy in institutional approaches to AI oversight, with most Member States opting to leverage existing regulatory bodies while only a few create new AI-specific institutions.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The unified approach has only been adopted by EU institutions through the European Data Protection Supervisor (EDPS).

“The unified approach, interestingly, has only been adopted by EU institutions through the European Data Protection Supervisor (EDPS).”

May combine two claims.

Luxembourg's implementation exemplifies the fragmented approach, with multiple sectoral authorities sharing market surveillance responsibilities while maintaining the National Commission for Data Protection as a horizontal coordinator.

“Luxembourg's implementation particularly exemplifies this approach, with multiple sectoral authorities sharing market surveillance responsibilities while maintaining the National Commission for Data Protection as a horizontal coordinator.”

May combine two claims.

Poland has opted for a primarily consolidated regulatory architecture with clearly defined jurisdictional boundaries.

“Poland has opted for a primarily consolidated regulatory architecture with clearly defined jurisdictional boundaries.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The centralized approach suggests a preference for strong governmental coordination in AI oversight, potentially facilitating faster decision-making and unified policy implementation.

“This approach suggests a preference for strong governmental coordination in AI oversight, potentially facilitating faster decision-making and unified policy implementation.”

May combine two claims.

Poland's approach reflects a strategic decision to create dedicated AI governance institutions rather than distributing responsibilities across existing regulatory bodies, enabling more focused expertise development while maintaining centralized coordination.

“This approach reflects Poland's strategic decision to create dedicated AI governance institutions rather than distributing responsibilities across existing regulatory bodies, enabling more focused expertise development while maintaining centralized coordination.”

Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

These divergent implementation strategies reflect fundamental differences in administrative traditions, institutional capacities, and conceptualizations of effective technology governance across the European regulatory landscape.

“These divergent implementation strategies reflect fundamental differences in administrative traditions, institutional capacities, and conceptualizations of effective technology governance across the European regulatory landscape.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The unified model combines both functions within a single independent authority, potentially offering clearer lines of accountability and more streamlined operations.

“This model combines both functions within a single independent authority, potentially offering clearer lines of accountability and more streamlined operations.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

It may also reflect practical considerations regarding the ambitious implementation timeline, as creating entirely new institutions by August 2025 could present significant organizational challenges.

“It may also reflect practical considerations regarding the ambitious implementation timeline, as creating entirely new institutions by August 2025 could present significant organizational challenges.”

May combine two claims.

While centralized approaches might offer more straightforward coordination, fragmented implementations might better address the multifaceted nature of AI regulation, from technical standards to market surveillance duties.

“While centralized approaches might offer more straightforward coordination, fragmented implementations might better address the multifaceted nature of AI regulation, from technical standards to market surveillance duties.”

May combine two claims.

The institutional continuity approach also presents potential challenges: existing bodies may need significant adaptation to handle AI-specific oversight responsibilities, and there might be risks of institutional path dependency (Chiodo, 2024), limiting the development of novel regulatory approaches needed for AI governance.

“However, the institutional continuity approach also presents potential challenges.Existing bodies may need significant adaptation to handle AI-specific oversight responsibilities, and there might be risks of institutional path dependency (Chiodo, 2024), limiting the development of novel regulatory approaches needed for AI governance.”

• cites Chiodo, 2024Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

The limited adoption of the unified approach by Member States might suggest practical challenges in implementing this approach at the national level, where existing regulatory landscapes are more complex, the market presents more AI systems for market surveillance activities, and robust capacity in AI is needed, notably towards the need for public acceptance (Middleton, 2022).

“However, its limited adoption by Member States might suggest practical challenges in implementing this approach at the national level, where existing regulatory landscapes are more complex, the market presents more AI systems for market surveillance activities, and robust capacity in AI is needed, notably towards the need for public acceptance (Middleton, 2022).”

• cites Middleton, 2022 sourceNot checked — no full text of the source

Classified as probably not a substantial claim. May combine two claims.

Hungary's creation of a dedicated AI Council and Poland's proposed Artificial Intelligence Development and Security Commission represent a more transformative approach to AI governance.

“Hungary's creation of a dedicated AI Council and Poland's proposed Artificial Intelligence Development and Security Commission represent a more transformative approach to AI governance.”

May combine two claims.

New AI-specific bodies might offer more flexibility in developing tailored oversight mechanisms, but they face the substantial challenge of building institutional capacity and expertise within a limited timeframe.

“Conversely, while new AI-specific bodies might offer more flexibility in developing tailored oversight mechanisms, they face the substantial challenge of building institutional capacity and expertise within a limited timeframe.”

May combine two claims.

Organizational fragmentation potentially enhances the depth of domain-specific understanding but simultaneously creates significant coordination burdens and risks of inconsistent implementation across regulatory domains.

“This organizational fragmentation potentially enhances the depth of domainspecific understanding but simultaneously creates significant coordination burdens and risks of inconsistent implementation across regulatory domains.”

May combine two claims.

These new institutions might offer advantages in terms of specialized focus and clear mandate, but they also face the challenge of building regulatory capacity and expertise from the ground up.

“These new institutions might offer advantages in terms of specialized focus and clear mandate, but they also face the challenge of building regulatory capacity and expertise from the ground up.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.

This institutional design dichotomy reflects deeper theoretical questions regarding the nature of effective technology governance - specifically, whether novel technological systems like artificial intelligence represent extensions of existing regulatory domains requiring evolutionary adaptation or fundamentally different governance challenges necessitating revolutionary institutional tran

“This institutional design dichotomy reflects deeper theoretical questions regarding the nature of effective technology governance -specifically, whether novel technological systems like artificial intelligence represent extensions of existing regulatory”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The choice to create new bodies might reflect these countries' vision of AI oversight as a distinct regulatory domain requiring dedicated institutional frameworks rather than an extension of existing regulatory functions, highlighting the tension in diverse AI public management and organization (Selten, 2024).

“The choice to create new bodies might reflect these countries' vision of AI oversight as a distinct regulatory domain requiring dedicated institutional frameworks rather than an extension of existing regulatory functions, highlighting the tension in diverse AI public management and organization (Selten, 2024).”

• cites Selten, 2024 sourceNot checked — no full text of the source

May combine two claims.

By distributing responsibilities across existing bodies, Member States can potentially meet the August 2025 deadline while leveraging established regulatory expertise.

“By distributing responsibilities across existing bodies, Member States can potentially meet the August 2025 deadline while leveraging established regulatory expertise.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Building on existing institutions might facilitate faster operational capability and more efficient resource utilization, particularly important given the comprehensive expertise requirements outlined in Article 70(3) of the AI Act.

“Moreover, building on existing institutions might facilitate faster operational capability and more efficient resource utilization, particularly important given the comprehensive expertise requirements outlined in Article 70(3) of the AI Act.”

• cites Article 70(3)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

The distributed approach operationalizes theoretical insights regarding the diversity of expertise necessary for effective technology governance, acknowledging that different regulatory domains require distinct forms of specialized knowledge that may not be readily consolidated within singular institutional structures.

“The distributed approach operationalizes theoretical insights regarding the diversity of expertise necessary for effective technology governance, acknowledging that different regulatory domains require distinct forms of specialized knowledge that may not be readily consolidated within singular institutional structures.”