Audrey Tang and Divya Siddarth on Outfitting Democracy for the AI Era
1Digest
This podcast discusses the exciting yet precarious role of AI in shaping democratic processes. The conversation centers around the experiences of Audrey Tang and Divya Siddharth, who have pioneered the use of AI to improve civic engagement, particularly in Taiwan's innovative approach to combating fraudulent investment schemes. Their work involved using AI to facilitate citizen deliberation through SMS outreach, random participant selection, and AI-moderated online discussions, resulting in citizen-generated solutions being rapidly implemented into law. The podcast highlights both the surprising successes—such as nuanced citizen participation and the discovery of "uncommon ground"—and the challenges—like crafting effective questions and shifting away from individualistic democratic models. Key themes include the three goals of democracy (buy-in, agency, good decisions), the importance of collective intelligence mechanisms (emphasizing the need to avoid homogenous outputs and untrusted agents), and the crucial distinction between collective intelligence and traditional polling. The discussion also touches upon the role of commercial centers in AI development, the ethical considerations of algorithmic transparency and human agency, and the integration of collective input into AI models. Ultimately, the podcast advocates for intelligently optimistic approaches to leveraging AI to improve government efficiency and effectiveness while maintaining trust and addressing potential biases.
Outlines
The Promise and Peril of AI in Democracy
The episode opens with the speakers' concerns and excitement about AI's potential impact on democracy, highlighting the surprising level of public trust in chatbots. The discussion sets the stage for exploring how to leverage this trust to build a more participatory and effective democratic system.
Audrey Tang and Divya Siddharth's Partnership
Introduces Audrey Tang and Divya Siddharth, highlighting their collaboration and experiences in using AI to improve civic engagement. Divya recounts how her work during the COVID-19 pandemic led her to Audrey's work in Taiwan, inspiring her to explore translating these successes to other contexts.
Taiwan's Response to Fraudulent Investment Schemes
Details Taiwan's innovative approach to combating fraudulent investment schemes using AI-facilitated citizen deliberation. This involved SMS outreach, random selection of participants, AI-moderated online discussions, and the rapid implementation of citizen-generated solutions into law.
Surprises and Challenges in AI-Facilitated Deliberative Democracy
The speakers discuss unexpected successes and challenges encountered while using AI in citizen deliberation processes. Positive surprises include the nuanced and engaged participation of citizens and the discovery of "uncommon ground." Challenges involve asking effective questions and shifting away from individualistic conceptions of democracy.
The Goals of Democracy and Collective Intelligence
Divya discusses the three key goals of democracy: buy-in and legitimacy, agency, and good decisions. She explains the trade-offs between these goals and how collective intelligence aims to preserve as many as possible while leveraging AI to improve decision-making processes.
Key Insights into Collective Intelligence
Audrey explains how removing reply and retweet buttons in online deliberation fosters more constructive and bridging conversations, leading to more coherent outcomes. The discussion shifts to misinformation and the importance of finding "uncommon ground" to depolarize discussions.
Collective Intelligence Mechanisms and AI
The conversation explores the importance of choosing appropriate collective intelligence mechanisms for different decision-making contexts. The speakers emphasize the need to avoid creating homogenous outputs, untrusted agents, and assigning autonomy to potentially harmful systems.
Collective Intelligence vs. Polling
The speakers differentiate collective intelligence from traditional polling, highlighting how group deliberation leads to more nuanced and creative solutions. They discuss the importance of considering how people's understanding of policies might change after implementation.
AI, Commercial Centers, and Societal Impact
The discussion turns to the role of commercial centers in AI development and the importance of societal considerations in the design and implementation of AI systems. The speakers emphasize the need to incorporate collective preferences into models and maintain human agency.
Integrating Collective Input into AI Models
Divya describes how collective input is integrated into AI models, using examples of working with Anthropic to revise model constitutions and gathering feedback from civil society organizations to improve model evaluations.
The Future of AI and Democracy
The conversation concludes with reflections on the future of AI and democracy, emphasizing the importance of trust in government, the need for intelligently optimistic approaches, and the potential of AI to improve government efficiency and effectiveness.
Keywords
Collective Intelligence
The synergistic process of combining knowledge and insights from multiple individuals for better outcomes than individual contributions. Involves structured deliberation and feedback.
Deliberative Democracy
A form of democracy emphasizing reasoned discussion and debate among citizens for collective decisions.
AI in Governance
The application of artificial intelligence to improve government efficiency, transparency, and citizen participation.
Citizen Deliberation
Processes where citizens engage in reasoned discussion and debate to shape policy decisions. Often facilitated by technology.
Bridging
Finding common ground and shared values between individuals or groups with differing viewpoints.
Uncommon Ground
Areas of agreement or shared values not immediately obvious, often discovered through deeper deliberation.
Algorithmic Transparency
The degree to which algorithms' processes and decision-making logic are understandable and accessible.
Participatory Democracy
A form of democracy emphasizing active citizen participation in decision-making.
Q&A
How can AI be used to improve civic participation and build a more participatory democracy?
AI facilitates citizen deliberation, enabling broader participation in policy decisions. Tools moderate discussions, summarize key points, and identify consensus areas, leading to more informed and inclusive decision-making.
What are some challenges in implementing AI-facilitated deliberative democracy?
Challenges include building trust, overcoming cynicism, ensuring accessibility and cultural appropriateness, and addressing algorithmic biases. A track record of success is crucial.
What are the ethical considerations surrounding AI in democratic processes?
Ethical considerations include algorithmic transparency, fairness, inclusivity, privacy protection, and preventing manipulation or misuse. Maintaining human agency and avoiding the erosion of core human values are paramount.
How can we ensure AI-facilitated democratic processes aren't dominated by special interests?
By designing systems prioritizing bridging and finding uncommon ground, rather than amplifying extreme viewpoints. Statistically representative sampling and careful moderation ensure a more balanced and inclusive process.
Show Notes
How can we take digitally-empowered democracy straight out of science fiction into reality?
This week, Reid and Aria are joined by two digital democracy pioneers, Audrey Tang and Divya Siddarth. Audrey Tang is a cyber ambassador-at-large and former inaugural Minister of Digital Affairs for Taiwan. Their pioneering initiatives like the Sunflower Movement, g0v, and vTaiwan have fought misinformation and influenced policy decisions. Audrey is also a senior research fellow for the Collective Intelligence Project, co-founded and directed by Divya Siddarth. Divya leads projects worldwide that give voice to the public in building better AI.
Audrey and Divya have spent years investigating how people converge on uncommon ground and how to beat polarization. They discuss the power of collective input and collaboration to create a cycle of collective intelligence. Plus, how AI tools—informed by public input—can enhance governance, digital life, and bring out the best in all of us.
For more info on the podcast and transcripts of all the episodes, visit https://www.possible.fm/podcast/
Topics:
3:07 - Hellos and intros
3:38 - Audrey and Divya's partnership backstory
7:06 - Challenges and surprises with AI in governance
10:19 - What problems do CIP's projects aim to solve?
13:04 - Breakthroughs in collective intelligence
15:08 - Fighting polarization in the age of social media
16:50 - Capitalizing on innate collective intelligence
18:51 - The difference between collective intelligence and democracy
20:33 - Design principles for democracy-forward AI
23:30 - Involving the public in AI model building
26:18 - Midroll
28:32 - Bringing Taiwan's civic wins worldwide
30:41 - Radical transparency in digital tools
35:18 - Why aren't Taiwan's initiatives landing in the U.S.
38:09 - Creating digital twins
39:52 - Reid and Aria's outlook on AI and democracy relations
42:04 - Who would Reid delegate AI centralization to?
45:05 - Rapid-fire
Select mentions:
Plurality by E. Glen Weyl, Audrey Tang, and Community
Global Dialogues — The Collective Intelligence Project
Alignment Assemblies — The Collective Intelligence Project
“Anthem” by Leonard Cohen
A Half Built Garden by Ruthanna Emrys
The Dispossessed: An Ambiguous Utopia by Ursula K. Le Guin
Terra Ignota series by Ada Palmer
Taiwan’s Digital Minister Knows How to Crush Covid-19: Trust | WIRED
How the Sunflower movement birthed a generation determined to protect Taiwan
Possible is an award-winning podcast that sketches out the brightest version of the future—and what it will take to get there. Most of all, it asks: what if, in the future, everything breaks humanity's way? Tune in for grounded and speculative takes on how technology—and, in particular, AI—is inspiring change and transforming the future. Hosted by Reid Hoffman and Aria Finger, each episode features an interview with an ambitious builder or deep thinker on a topic, from art to geopolitics and from healthcare to education. These conversations also showcase another kind of guest: AI. Each episode seeks to enhance and advance our discussion about what humanity could possibly get right if we leverage technology—and our collective effort—effectively.
Table of contents
United States
