Publication 6 February 2025

AI Governance: Empowering Civil Society

The final report of the AI Dialogues series

AUTHORS & CONTRIBUTORS

  • Théophile LENOIR, Head Reporteur of the AI Dialogues

  • Jessica GALISSAIRE, Studies & Partnerships Manager, Renaissance Numérique

  • Jean-François LUCAS, General Manager, Renaissance Numérique

This publication is the result of the three "AI Dialogues" organised by Renaissance Numérique over the course of 2024. These full-day high-level private seminars gathered more than 40 experts from academia, civil society, policy and industry, to discuss key issues related to the global governance of AI. Following the publication of two interim notes, "Is international AI governance achievable ?" and "The sites & actors of AI governance", this final report examines the role of civil society in the governance of AI. It explores both the current state of participation and the barriers that civil society organisations face. Ahead of the Paris AI Action Summit, it proposes three concrete measures for a more meaningful inclusion of civil society actors in global AI governance efforts.

Context

The meaningful participation of civil society is a challenge in global governance initiatives. In the context of artificial intelligence (AI), organisations and groups such as the G7, the G20, the Council of Europe, the Organisation for Economic Co-operation and Development (OECD), and the Global Partnership on Artificial Intelligence (GPAI) – which merged with the OECD in 2024 – have all sought to participate in a global effort to build a common understanding of the transformations induced by artificial intelligence. Most recently, the United Nations presented its strategy on how to enhance global cooperation around the governance of artificial intelligence. In these efforts, the voice of civil society is often sought after, but questions remain regarding the extent to which it is accounted for. The preoccupation regarding the meaningful inclusion of civil society representatives in multi-stakeholder governance processes is not new, and continues to be an important source of debates unfolding in the background of international governance discussions.

Virtually all parties designing policy and governance initiatives claim to rely on inputs from civil society organisations. But including civil society in high-level decision making is difficult. How to identify relevant contributions amongst thousands of sollicitations? Are civil society organisations truly representative? At which stages of decision-making are they most useful? Do they need a voting right in some instances for their participation to be truly meaningful? How to make the best use of their limited availability when tens of similar initiatives run in parallel? It is rarely clear how international or national institutions address these difficult questions. Few provide clear information about the processes through which civil society is recruited and participates in governance, for example.

Structure

Against this backdrop, this report examines the role of civil society in the governance of AI. It explores both the current state of participation and the barriers that civil society organisations face. It is structured as follows. The first section provides general considerations about the participation of civil society in the global governance of AI: what is the global governance of AI, why involving civil society in efforts to shape it matters, and concrete examples of how civil society currently takes part in those efforts. The second section highlights key challenges associated with civil society’s involvement in the global governance of AI, such as risks of tokenisation, geographic representation imbalances, lack of funding and time, and increasing need for coordination. To go beyond this status quo, the report concludes on a call to action ahead of the upcoming AI Action Summit that will be held in Paris on February 10 and 11, 2025.

Global AI Governance: Our Call to Action

Funding limitations, lack of expertise or time, coordination difficulties, and risks of tokenisation, limit civil society’s ability to fully engage in global AI governance efforts. Addressing these barriers is necessary to ensure civil society can advocate effectively for a responsible AI ecosystem.

Looking forward, strengthening civil society’s influence in AI governance will require sustained commitment from governments, international organisations, and civil society itself. In the run-up to the Paris AI Action Summit, we request organisers to urgently implement three concrete measures to strengthen and facilitate the involvement of civil society in global AI governance, during and especially after the summit:

1 - Support more open and transparent selection processes

Clarifying and opening the selection processes for civil society’s participation in expert groups (in standardisation processes, within the European commission, etc.) and making them public, is essential to ensure fair representation and restore trust. Opportunities for civil society participation should also be widened: we recommend opening up the participation spaces organised by public authorities or private players to a wider spectrum of organisations, beyond the most established ones.

2 - Impose a duty to respond to contributions

There are often no mechanisms in place to ensure that civil society contributions are actually taken into account. We call on the summit participants to encourage the establishment of clear mechanisms for feedback and reporting on the integration of civil society inputs into AI-related policies and practices. These could be inspired by the OECD Guidelines for Citizen Participation Processes, which provide useful recommendations and concrete examples on how to “close the feedback loop”.

3 - Facilitate funding

Many civil society organisations have extremely limited resources. Setting up dedicated funds or processes to encourage their funding, would enable them to contribute effectively and sustainably to AI governance efforts. In connection with recommendation 1, it would also be useful to restructure the funding models of working groups, committees, technical groups, etc. to make sure funding is allocated to support the inclusion of civil society in these fora, in a transparent and independent manner.

By creating these conditions, civil society can contribute more meaningfully to shaping AI policies that protect public interests and address both immediate and long term risks associated with AI technologies. Ultimately, a robust and representative civil society presence in AI governance will be instrumental in developing AI systems that reflect collective values and serve the public good.

Simply consulting civil society in a superficial manner or highlighting its presence through media operations is no longer enough. It is now time to give civil society representatives a full and permanent seat at the table, so that they can become leading partners in the development of policies and guidelines, and in adjusting AI tools. The Paris AI Action Summit provides an opportunity to rebalance the powers between all the players who should have a say in the global governance of AI. Let’s hope it won’t be a missed opportunity.

Methodology

In early 2024, Renaissance Numérique launched a three-day cycle of high-level seminars focusing on the global governance of Artificial Intelligence: the AI Dialogues. Inspired by a previous successful series, the Metaverse Dialogues, this project aimed at bringing together European and international experts to discuss international, European, national and local governance issues related to artificial intelligence (AI). The first Dialogue, titled “Is International AI Governance Achievable?”, took place in Geneva on 26 April, 2024. The second one, “The Multiple Actors of International AI Governance”, took place in Brussels on 27 June, 2024. The last one, “How to involve citizens & civil society in the global governance of AI?”, which greatly inspired this report, took place in Paris on 10 October, 2024.

To complement the input gathered during the Dialogues, the authors conducted extensive desk research.

Academic & Funding Partners

This project was made possible thanks to Renaissance Numérique’s Academic and Funding Partners. We warmly thank them for their support.

More About the AI Dialogues

To learn more about the AI Dialogues series of events organised by Renaissance Numérique, and the process that led to the publication of this report, visit the AI Dialogues website at: www.ai-dialogues.org/

Join the movement

To complement this report, Renaissance Numérique has initiated an opinion piece based on its call to action, calling for a more meaningful inclusion of civil society in global AI governance efforts.  More than 40 civil society actors, academics and experts from over the world have signed onto it, and the piece will remain open for signatures. Interested in learning more or joining the movement? Click here!

More on this subject