Presentation

The European Union (EU) has developed a new and unique legislative framework for Artificial Intelligence (AI) systems put in place on the internal market. Regulation 2024/1689, called “Artificial Intelligence Act”, entered into force on 1 August 2024 and will be progressively applicable between 2 February 2025 and 2 August 2027 depending on the provisions concerned. The text is underpinned by a dual rationale. On the one hand, it aims to ensures the free movement of AI-based goods and services while supporting innovation and economic growth in the EU. On the other hand, it seeks to promote trustworthy AI systems, guaranteeing the protection of health, safety and fundamental rights against harmful effects these systems may have on people and on society.

The AI Act provides for three main categories of legal provisions : first, a list of prohibited AI practices; second, harmonised rules applicable to marketed AI systems, following a risk-based approach (including provisions on innovation and on general purpose AI models); and third, a comprehensive public enforcement scheme. It consists of more than 100 articles, 180 recitals and 13 annexes. It is thus a massive and complex regulatory framework that both public and private organisations – dealing with AI systems and active on the EU market – will have to master and implement in the upcoming months and years. Therefore, it will be crucial for AI industry and AI practioners, including public authorities, to set up an action plan to comply with the AI Act.

The AAIAC workshop offers a two-days’ workshop exploring ways and methodologies to achieve the EU AI Act compliance. The objective is to analyse the new legal framework according to a “compliance mode” aimed at professional operators and interested parties (such as civil society representatives) of the AI ecosystem. The workshop is conceived as a platform for sharing legal knowledge and compliance practices, based on both academic and practical analysis.

The methodology adopted is inspired by ISO standards on management systems and its continuum in the field of AI. One of the main goals of implementing a compliance management system for an organisation is to be able to demonstrate its commitments to comply with relevant normative frameworks, whether based on hard or soft law, including best practices and ethics. Transposed to the field of AI, management system’s standards offers a “structured way [for organisations] to manage risks and opportunities associated with AI, balancing innovation with governance”. This standard rationale takes on a totally new dimension in the forthcoming EU AI regulatory context. It is therefore essential to explore how the AI management system of organisations should be adapted in practice to the AI Act’s requirements and obligations. The workshop will open discussions on explaining, commenting on and questioning the AI Act with a view to its implementation by AI operators, including within their management system.

Program

Thursday 13 February

Opening

14:00 : How to Achieve Compliance with the EU AI Act ? Issues and Methods
Marion Ho-Dac, Artois Univ. & Cécile Pellegrini, UCLy

Part 1 - Organising the AI act compliance

Panel 1 - Establishing the Scope of the Compliance Scheme

Chair : Michel Cannarsa (UCLy)

14.15 : AI Systems & Models Taxonomy under the AI Act
Juliette Sénéchal, Lille Univ.

AI Operators under the AI Act
Marco Pasqua, Milan Catholic Univ.

AI Legislative Frameworks coordinated with the AI Act
Beatrice Schütte, Helsinki Univ.

Q&A

Break

Panel 2 - Mapping the Major High-Risk AI Systems Requirements

Chair : Mathieu Guillermin, UCLy

15h45 : Risk Management System under the AI Act
Amélie Favreau, Grenoble Univ.

Data Governance and Management Practices under the AI Act
Jean-Marc Van Gyseghem, Namur Univ.

Transparency (and Related Requirements) under the AI Act
Florence Guillaume, Neuchâtel Univ.

Support, Competence, Training, Procedures, Documentation in Organisations
Nathalie Nevejans, Artois Univ.

Q&A

17.30 : Keynote - EU AI Act Compliance Challenge from an Organisation Perspective
Axel Cypel, Head of AI projects & author

18.30 : End of 1rst day

Cocktail Reception

Friday 14 February

Part 2 - Implementing the AI act compliance

Panel 3 - Compliance Tools and Processes

Chair : Franck Violet, UCLy

9:30 : European harmonised Standards under the AI Act

- Interplay/Coordination with International Standards
Olia Kanevskaia Whitaker, Utrecht Univ.

- AI Trustworthiness Characteristics
Piercosma Bisconti, National Interuniversity Consortium for Computer Sciences

EU Commission and other Authorities Guidelines, Delegated Acts, Code of Practice
Céline Castets-Renard, Ottawa Univ.

Q&A

10.30 : Break

Specific Compliance Schemes under the AI Act

Chair : Alexandre Palanco, UCLy

11.00 : Deployers of FRIA (Fundamental Rights Impact Assessment)
Lamprini Xenou, Paris Est Créteil Univ.

Regulatory Sandboxes under the AI Act
Gaurav Sharma, International AI Policy and Advocacy Advisor

GPAI (General Purpose AI) Transparency and Systemic Risks Management
Arnaud Latil, Paris-Sorbonne Univ.

Q&A

12:30 : Lunch break

Panel 4 - Compliance Evaluation and Enforcement

Chair : Gäelle Marti, Lyon 3 Univ.

14.00 : Conformity Assessment Procedure (including Audit and Monitoring)
Bernard Guillaume, LNE

Cooperation and Control by Public Authorities (including EU AI Office)
AI Office representative

Use Cases Round-Table : AI Act Compliance Action Plan in Practice
With :
Yves Raisin, BioMérieux - Capucine Hurel, Dalloz-Lefebvre & Hugo Ruggieri, Doctrine.fr

Q&A

16:30 : Closing

Open to all with registration

Registration : https://www.ucly.fr/l-ucly/agenda/workshop-ai-act-compliance/

Formation éligible pour les avocats français au titre de la formation continue

- Tarif : 50 € - Validation de 4h par demi-journée

Contact : ndagron@univ-catholyon.fr