Event Details
On 29 November 2024, CDL hosted Assoc Prof Gilad Abiri from Peking University School of Transnational Law for an internal research seminar on Public Constitutional AI.
As artificial intelligence (AI) systems become increasingly powerful and pervasive, influencing decisions across domains from markets to media, governance to healthcare, ensuring their legitimacy emerges as a critical challenge. Assoc Prof Abiri proposed Public Constitutional AI as a regulatory framework for addressing this challenge by directly involving citizens in the development of the principles and values that guide AI systems.
Drawing on the concept of Constitutional AI, which seeks to hardcode explicit principles into AI systems to make their decision-making more transparent and accountable, he argued that for such an approach to truly secure the legitimacy of AI governance, it must be grounded in public participation and deliberation. Rather than relying on constitutions developed by private corporations, Public Constitutional AI engages citizens in drafting the principles that govern AI systems operating in their jurisdictions.
By making the AI constitution a product of democratic process, Public Constitutional AI offers a path to making these often opaque and abstract systems more legible and accountable to the communities they serve. We thank Assoc Prof Abiri for highlighting that the legitimacy of AI governance depends not just on the content of the principles embedded in these systems, but on the process by which those principles are developed and the extent to which they reflect the shared values and will of the public.
