Global citizen deliberation: a way to achieve better AI governance?

This recap was written by Lina Srivastava, Founder, Center for Transformational Change who guest moderated this Salon.

The governance of artificial intelligence—covering its design, deployment, and use—is rapidly emerging as one of the most critical social and political issues of our time. Our political landscape is marked by a polycrisis of rising violence, threats to democracy, escalating inequality, and climate collapse. Profit-driven AI solutions often worsen these issues, forcing communities to confront AI’s impact amid a deteriorating information environment.

As AI increasingly influences daily life, from decision-making to economic structures, the need for effective governance grows more urgent. Ethical oversight and accountability are crucial to safeguarding individual and community rights.

While the development of advanced AI tools remains concentrated in a few wealthy Western countries, its effects are felt worldwide. As AI reshapes supply chains, labor markets, agriculture, and media, its governance must be inclusive and representative, and voices from a diversity of regions, cultures, and sectors should be integrated into governance structures. 

Global citizen deliberation: a tool to improve AI Governance?

Global citizen deliberation may offer one solution to this challenge. Democratic engagement methodologies, such as citizen assemblies and community level discussions, bring representative groups together to deliberate on pressing issues with balanced expert testimony. These efforts can effectively shape public debate and decision-making.

At our September 20, 2024 Technology Salon NYC we explored the question: “How Can Global Public Deliberation Shape AI?” We had four expert discussants and a guest moderator (listed below), joined by over 40 participants navigating this evolving landscape.

• Tim Davies, Director of Research and Practice at Connected by Data,
• Helene Landemore, Professor of Political Science, Yale University & Fellow of Ethics in AI Institute, University of Oxford,
• Richard Wilson, ISWE Foundation / Coalition for a Global Citizens Assembly
• Kiito Shilongo, Senior Technology Policy Researcher
• Lina Srivastava, Founder, Center for Transformational Change (guest moderator).

The Salon, sponsored by Connected by Data and ISWE Foundation and hosted by Thoughtworks, also functioned as the launch of Global Citizen Deliberation on Artificial Intelligence: Options and design considerations, a report co-authored by Connected by Data and ISWE that underpinned the conversation. 

Over the course of the Salon, participants and the discussants tackled a wide range of critical questions regarding the role of global citizen deliberation in shaping the future of artificial intelligence. They explored why such deliberation may be essential and how lessons from existing practices in AI and related fields, like climate governance, could inform their approach. 

They considered what key questions could be posed to global publics, and what forms these assemblies might take. This part of the discussion included a deep dive on the risks of participation-washing and embedded power imbalances, specifically regarding who should be involved in setting the agenda and participating in the assembly.

Additionally, discussants identified opportunities to advance the design of these practices, ensuring that the deliberative process is both inclusive and effective in addressing the complexities of AI governance.

Defining the Scope 

To ground the conversation, one lead discussant began by establishing the room’s shared understanding of AI and public deliberation. 

AI was defined broadly to encompass not only recent advancements like ChatGPT but also, in recognition of long-standing ethical challenges, technologies that have been in use for much longer. 

Public deliberation was characterized by several key principles, emphasizing that effective deliberation goes beyond opinion gathering; it requires informed contributions and sufficient time for reflection, as well as:

• Intentional diversity: Actively recruiting a range of voices to ensure representation from various backgrounds. 
• Expert input: Providing balanced insights from experts to support informed discussions. 
• Facilitative dialogue: Encouraging conversations that foster shared understanding and bridge gaps. 
• Action-oriented outcomes: Ensuring that deliberations lead to tangible policy changes and implementation.

The Need, Challenges, and Opportunities of Global Deliberation

One proposed approach for global public deliberation identified four process stages:

• Diagnosing
• Prototyping
• Implementing 
• Establishing.  

The discussant suggested that currently, the process to establish global public deliberation is in the diagnostic stage, facing challenges in facilitating multilingual and asynchronous deliberations on a global scale.

One lead discussant did note two positive trends emerging among the general public: (1) increased knowledge: More people are gaining a deeper understanding of AI issues in recent years; and (2) heightened interest: There is a greater public willingness to engage in discussions about AI’s impact. But current AI governance frameworks still exhibit significant gaps in:

• Geographic representation: A lack of comprehensive global participation.
• Multiple perspectives: Insufficient inclusion of voices with lived experiences, especially those most affected by AI.
• Stakeholder engagement: The need for more input from those directly impacted by AI-related decisions.

To address these gaps, a global citizen’s assembly or worldwide deliberation on AI was proposed. Key considerations for this initiative include:

• Topics for discussion: Potential areas might encompass visions and values for AI, governance regulations, and specific applications.
• Format of deliberation: Options range from smaller gatherings of global representatives to large-scale online engagement.
• Advocacy strategies: Developing effective methods for promoting change and implementing decisions was emphasized.

Power Dynamics and Theories of Change

Participants noted that while the methodologies of creating both a culture and a practice of global public deliberation were being mapped out, the question of power imbalances in participation, inclusion, and decision-making persist. A significant part of the Salon discussion was dedicated to the intricate power dynamics shaping global dialogues, with examples raised from prior efforts to create external deliberative bodies that were akin to participation-washing. 
To address these risks, participants and discussants explored the following theories and practices:

Social Movements: These were identified as indispensable for challenging prevailing power dynamics and elevating a spectrum of perspectives. By weaving together these strategies, we can foster inclusive deliberation that bridges global principles with local realities, creating a truly representative dialogue. A few participants versed in social movements and advocacy emphasized the urgency of a robust and multifaceted approach that unites both “insider” and “outsider” strategies:

• Insider: Cultivating relationships with policymakers and actively engaging with established institutions to navigate and influence existing policy frameworks is essential.
• Outsider: Empowering social movements and marginalized communities is critical for exerting pressure outside conventional power structures, ensuring that diverse voices are not just heard but amplified.

Lived Experience: It was noted that elevating the voices of those who bear the direct impact of AI and creating avenues for active listening is imperative. Rallying communities around abstract concepts like privacy may pose challenges, but the increasingly tangible effects of AI on employment and livelihoods are seen as likely to galvanize broader public engagement. As these realities unfold, they may serve as powerful catalysts for collective action and advocacy, driving the urgent demand for responsible AI governance.

Balancing Global and Local Governance: Another recurring theme was the tension between establishing global norms and implementing them at the local level. Governance requires setting international standards, while regulation often reflects individual nations’ interests, leading to inconsistencies. Immediate efforts should focus on creating global norms around unacceptable uses of AI, akin to treaties on nuclear or biological weapons. Social movements can play a crucial role in advocating for these standards and influencing major powers to codify and enforce them.

The Importance of Narrative and Vision: Participants emphasized the need to shift from solely addressing AI’s harms to articulating positive visions for its potential. Mobilizing around opportunities—such as feminist AI or AI for social good—can inspire proactive engagement and collective narratives that emphasize the transformative benefits of AI, rather than merely its risks.

Institutional Reform and Engagement: The need for institutional reform and greater inclusivity in AI governance was highlighted. Engaging diverse stakeholders, including those from sectors traditionally excluded from AI discussions, is essential. Recent shifts towards more open dialogues in spaces like trade indicate a growing responsibility to democratize decision-making processes that were previously dominated by corporate interests. 

Technical Challenges and Collaboration: The technical complexities of AI governance were a focal point, particularly regarding the vast scale and intricacy of modern AI systems. One participant noted that major tech companies often operate through extensive networks, while human rights organizations typically lack the resources to compete. It was suggested that effective governance may thus require collaboration with these companies, as well as innovative solutions to develop inclusive, multilingual, and contextually appropriate technical infrastructure, especially in regions with diverse languages.

A Call to Action

Concrete actions to advance global public deliberation on AI governance were laid out:

• Establish clear benchmarks: Develop collaborative standards to measure AI’s social impact, safety, and fairness.
• Expand public education: Increase awareness and understanding of AI implications through accessible resources and dialogues.
• Create safe spaces: Intentionally include marginalized voices in deliberation processes, ensuring their experiences shape governance discussions.
• Re-imagine governance: Balance insider and outsider strategies to foster inclusive decision-making.
• Bridge gaps between stakeholders: Connect technologists with human rights advocates to create robust AI governance models.
• Mobilize social movements: Utilize grassroots efforts to push for global norms and influence national regulations.
• Articulate a positive vision: Emphasize AI’s potential benefits to foster proactive engagement.
• Strengthen collaborative networks: Facilitate platforms for coordinated action among communities, policymakers, and movement builders.

Navigating the complexities of AI governance requires intentional and nuanced approaches that acknowledge the challenges of true participation in global public deliberation. The Salon identified clear steps toward creating an inclusive framework, emphasizing collaboration, diverse perspectives, and a blend of insider and outsider tactics. By managing these complexities effectively, we can shape a more equitable and responsive governance landscape for AI.

---

Technology Salons run under Chatham House Rule, so no attribution has been made in this post. If you’d like to join us for a Salon, sign up here. If you’d like to suggest a topic please get in touch!

We are looking for sponsorship to help cover the costs of preparing and hosting salons – please contact Linda if you would like to discuss financial support for Tech Salon.