Students sit around a table in a room with other tables and students talking
Hélène Landemore believes in participatory democracy — possibly with some positive role for artificial intelligence.
So, as government leaders, academics, and representatives from industry and civil society from more than 100 countries gathered in Paris last month for the Artificial Intelligence Action Summit, Landemore — with MacMillan Center for International and Area Studies at Yale associate research fellow Théophile Pénigaud, and ISPS postdoctoral fellow Antonin Lacelle-Webster — helped convene a parallel student assembly on AI.
The event also served to help launch a project for a global deliberative process on AI that emerged from a conference Landemore organized last year through Democratic Innovations, a program at Yale’s Institution for Social and Policy Studies (ISPS) designed to identify and test new ideas for improving the quality of democratic representation and governance.
A collaborative effort among ISPS, the MacMillan Center, the École Normale Supérieure (ENS), Missions Publiques, and the AI-2050 program at Schmidt Sciences, the assembly engaged a random selection of 70 college students from within the ENS-PSL network of different ages and disciplines, including sociology, biology, philosophy, neuroscience, economics, management, computer science, mathematics, and more.
The students consulted internationally recognized experts on AI to debate risks and opportunities of the technology at both national and global levels, as well as within their own institutions.
“Since they were holding this giant conference on AI in Paris, ENS Director Frédéric Worms and I thought we could try a citizen’s assembly on the sidelines,” said Landemore, a professor of political science who helps lead ISPS’s Democratic Innovations program. “That way we could get students involved in an active role, allow them to practice democratic deliberation instead of lecturing to them about it, and perhaps insert this assembly into our larger, long-term global project on inclusiveness in AI.”
Audrey Tang gestures as she speaks to students seated at a table
ISPS Director Alan Gerber, Sterling Professor of Political Science, expressed excitement at how connections established at last year’s Democratic Innovations conference in New Haven continue to reverberate.
At Yale, when we are doing our job well, we can produce these sparks of inspiration — stimulate thought and action,” Gerber said. “I am always thrilled to see how discussions begun here can grow into something substantial on a local, national, or even global scale.”
Facilitated by professionals from Missions Publiques and students trained by the organization, the assembly featured an AI prototype to synthesize input manually entered by students and groups at various stages of the deliberations and link summaries back to the original discussion with a search function.
“This raises the question of what role such a tool might — or might not — play in the future of deliberative and participatory democracy,” Landemore said.
On the first morning, students shared their AI experiences, set discussion rules, and shared AI’s impact on their lives. In the afternoon, they considered AI’s risks and opportunities and prepared questions for experts.
The next day, they met in small groups with experts in a speed dating format to explore democratic processes, ethics, and philosophy before deliberating and presenting their recommendations for ENS-PSL and beyond. Among international figures, experts included Rob Reich, McGregor-Girand Professor of Social Ethics of Science and Technology at Stanford University, Audrey Tang, at-large cyber ambassador and the first digital minister of Taiwan; Annette Zimmermann, assistant professor of philosophy at the University of Wisconsin-Madison; and Marc-Antoine Dilhac, associate professor of philosophy at the University of Montreal.
The students recommended creating guidelines and criteria for AI at ENS-PSL through discussions with faculty and students, developing proprietary AI capabilities at the school to ensure control and unify their use, holding mandatory training sessions, establishing a laboratory dedicating to studying large language models focused on understanding and not merely boosting performance, and building an information platform dedicated to explaining AI usage.
On a global scale, the students recommended developing a sovereign AI for Europe with strong democratic oversight, transparency, and independence; incorporating boundaries into AI usage; establishing regulations to constrain AI integration into certain tasks; encouraging companies to reduce energy consumption; involving citizens in drafting laws and regulations related to AI; and promoting the financialization of AI by providing access to markets and ensuring that bond holders are represented in government.
At the end of the last day, the students were also invited to join a 90-minute private meeting with French Prime Minister François Bayrou. Right after the meeting, Bayrou delivered a public speech to over 200 people, which he concluded on a reference to “participatory democracy.” Landemore expressed her pleasure with the success and seeming impact of the student assembly.
A table in a conference room with students and the prime minister of France
“This shows to me there is an appetite for this,” she said. “It’s not what the political world is primarily interested in. But once you launch these things, people are excited. They want to learn more about it and join in.”
She felt particularly pleased to involve students alongside a major international summit that had little room for the average citizen and to plant a seed for systemic change.
“The conversation on AI has mostly been a debate between technologists and government representatives, but what about the regular folks?” Landemore said. “I told the students they were not there to learn primarily, as students, but to take ownership of this process, as citizens. We need to change the attitude. Show them a different, more active way of doing democracy.”