In their groundbreaking report, Philippe PRADAL of Alpes-Maritimes and Stéphane RAMBAUD of Var presented vital recommendations to the Law Commission of the French Parliament concerning generative artificial intelligence (AGI). These recommendations aim to effectively regulate AGI development and usage by the State, companies, and individuals while fostering innovation in France’s thriving AI sector.

The crucial recommendations are as follows:

1. Appoint an AGI ambassador to oversee international discussions on regulating AGI.
2. Facilitate knowledge and experience sharing on AI control at a European level, specifically focusing on generative control.
3. Adapt AI usage constraints depending on whether it is employed by Member States or private actors.
4. Prioritize audience analysis of AGIs before determining risk levels to encourage experimentation and innovation without penalizing newcomers.
5. Mandate compliance with GDPR and technology-neutral principles for AGI services within the European Union.
6. Encourage soft law approaches and compliance mechanisms alongside binding measures.
7. Allow EU Member States flexibility in supporting innovation.
8. Promote intergovernmental cooperation among voluntary European countries to establish an “Airbus of AGI,” redirecting corporate funds towards AI innovation in Member States.
9. Transform CNIL into a High Authority responsible for data protection and generative AI control.
10. Ensure coordination between this new authority and Arcom, which regulates platforms.
11. Foster the regulator’s role as a guide in navigating AGI challenges.
12. Identify public sector tasks that could potentially be delegated wholly or partially to generative AI systems.
13. Provide training for public officials regarding opportunities and risks associated with using AGI systems.
14. Encourage public buyers to adopt trusted and labeled AGI systems.
15. Establish a multi-year strategy for generative AI, dedicating specific components to its implementation in administrations.
16. Develop internal expertise in AGI application development based on AGI models.
17. Create a permanent forum for dialogue between administrations regarding AGI usage.
18. Aim to provide the public with a reliable conversational agent based on generative AI by 2027, enhancing access to law and administrative information.
19. Promote labeling of generative AI-generated content for election propaganda purposes.
20. Extend the requirement for information provision when interacting with AGI systems to chatbots utilized in election campaigns.
21. Establish a suitable legal framework for experimenting with generative AI systems in the sovereign domain.
22. Provide financial support through grants and computing resources to French researchers in AGI field.
23. Amend Article 132-79 of the Criminal Code to include content generated by an AGI under aggravating circumstances related to cryptographic means’ use.
24. Penalize deepfakes created without consent from depicted individuals.
25. Initiate discussions on counterfeiting offenses, addressing the practice of masking pre-existing content reproduction using AGI assistance.
26. Introduce measures for fraudulent distortion of truth by claiming authorship of writings or media substantially produced by an IAG (Intelligent Artificial General) system during forgery cases.
27. Adapt Article 323-1 of the Criminal Code to punish poisoning of an AGI’s training data or clandestine modification of its algorithm.
28. Enhance public authorities’ capability to identify artificial intelligence-generated content and prevent misuse of AGIs.
29. Modify liability regulations for AGIs, taking their specific characteristics into account and reducing burden of proof to level the playing field between users and suppliers.
30. Consider liability issues for service providers who rely on externally designed AGIs.
31. Reform class action laws pertaining to personal data protection, expanding standing and scope to encompass all damages associated with an AGI system.
32. Establish specialized jurisdictions dedicated to handling AGI litigation cases efficiently.
33. Entrust OPECST with continuous monitoring of artificial intelligence-related issues.

These recommendations offer a comprehensive framework for regulating AGI while encouraging innovation, strengthening data protection measures, and ensuring responsible usage in various sectors.