Background and Context
Taiwan's democracy is grappling with AI, geopolitics, and social divisions. This has sparked calls for participatory engagement via AI tools like Polis to strengthen institutions. MODA has commissioned CIDS to lead "Alignment Assemblies," seeking to integrate human-centered perspectives into the nation’s AI governance. By leveraging data from Polis.tw and hosting these specialized deliberative sessions, CIDS gathers insights from professional communities and the public to identify critical debates surrounding Generative AI in Taiwan. This collaboration aims to understand localized concerns, social needs, and assist the government in crafting robust policies and regulatory tools that align Taiwan’s AI development with global trends.
By hosting two public discussion sessions, we are gathering a wide range of opinions to shape specific AI policies and understand their social effects. This inclusive approach involves both individuals and stakeholders, ensuring that professional private-sector voices are heard. It is a prime example of public-private collaboration that guarantees representation, providing a legitimate and essential basis for the government’s future actions.
CIDS x MODA collaboration in Governance and AI
With the rising global trend of integrating and developing AI in governance, the Taiwan Ministry of Digital Affairs (MODA) has commissioned the Centre for Innovative Democracy and Sustainability (CIDS) to lead "Alignment Assemblies," seeking to integrate human-centered perspectives into AI governance capacity.
By leveraging data from Polis.tw and hosting these specialized deliberative sessions, CIDS gathers insights from professional communities and the public to identify critical debates surrounding Generative AI in Taiwan. This collaboration aims to understand localized concerns and social needs, assisting the government in crafting policies and regulatory tools to align Taiwan’s AI development with global trends.
This inclusive approach involves both individuals and stakeholders, ensuring that private-sector voices are heard. This is a prime example of public-private collaboration that guarantees representation by providing a legitimate and essential basis for the government’s future actions.
Organizing Actors and Resources
Lead organizations
The Centre for Innovative Democracy and Sustainability (CIDS) leverages innovative thinking to forge new pathways for public participation in policy-making. Our research focuses on the multifaceted dimensions of technology governance, including human rights in relation to NIMBY (Not In My Back Yard) facilities, social impact assessments of technological innovation, and public engagement in nuclear waste repository siting. By bringing critical public issues to the forefront, CIDS aims to synthesize diverse perspectives through multi-stakeholder forums to develop the most suitable solutions for societal development.
Governmental involvement
The Ministry of Digital Affairs (MODA) seeks to utilize "Alignment Assemblies" on the principles of deliberative democracy. These assemblies aim to explore and understand the visions, policy preferences, and concerns of both professional communities and the general public regarding Generative AI. This initiative will serve as a foundation for aligning Taiwan's localized key AI issues and policy guidelines with global trends.
Professional roles (facilitators, experts, planners, NGOs)
The participants primarily consist of community members with professional expertise or backgrounds in emerging technologies. Consequently, throughout the deliberative process, they offer insights grounded in a stronger technical foundation in contrast to the general public, enabling a more focused and substantial advancement of the issues.
Participatory Design and Methods
Through two sessions of public deliberation, this initiative broadly solicits diverse perspectives to shape concrete policies and define the social impact of AI. By engaging individuals and key stakeholder communities, the project ensures that professional insights from the private sector are fully integrated. This process serves as a model for public-private partnership (PPP), establishing a robust foundation of legitimacy and a critical reference for national policy-making.
Furthermore, by analyzing the current landscape of Generative AI alongside data from Polis.tw, we identify the primary scenarios envisioned by the public. Finally, through the strategic design of these two deliberative workshops, we explore the underlying causes of various points of contention, providing a comprehensive evidence-based reference for future policy planning.
Phases
Phase 1 : Immersive Skit
As part of the session's orientation, we utilize an "Immersive Skit" to contextualize complex AI issues. This method reduces cognitive load and ensures that participants, regardless of their prior knowledge, can quickly align with the deliberative goals of the assembly both logically and emotionally.
Phase 2 : Grouping and Agenda Setting
Participants will be divided into six thematic groups for in-depth deliberation. The discussion tracks include: Government Governance (Internal Application), Emerging Legal Frameworks (Regulatory Mechanisms), Information Identification, Data Openness, Intellectual Property, and Education and Cognitive Impact.
Phase 3 : Polis Data Analysis
Prior to the workshop design deadline (as of August 1, 2023), the polis.tw survey engaged a total of 403 participants. The platform recorded 8,591 votes, with an average of over 21 votes per user. Participants contributed 126 unique statements, and 260 individuals were categorized into specific opinion clusters via the platform’s algorithm. These preference patterns regarding Generative AI applications were analyzed and grouped into thematic tracks, which were then integrated into the deliberative workshop sessions as a foundational reference for participant discussion.
Phase 4 : Discussion Model and Toolkit Design
The discussion is structured as an open-ended focus group tailored for informed public. Unlike broad consultations, this model uses case studies across themes. Table Leads (facilitators) synthesize views and keep things on track. It taps participants' expertise to turn discussions into policy insights.
Phase 5 : The Deliberative Framework consists of two primary parts
5.1 Warm-up & Narrative Connection
This phase initiates discussion through guiding questions designed to help participants connect their personal experiences with the core issues. By fostering mutual understanding within each group, this stage prepares participants for deeper analysis. Each session concludes with cross-group sharing to help participants build a holistic vision of the AI landscape.
5.2 Policy Synthesis & Action Planning
Building on the cross-group insights, facilitators guide participants through two fundamental policy frameworks: Top-down Public Governance and Bottom-up Civic Engagement. Groups then identify policy tools to address the risks and challenges identified earlier. By prioritizing actions and combining strategic conditions, participants co-create innovative, executable action plans for a future where humanity and AI coexist harmoniously.
Tools used
Polis
Pol.is is a real-time survey system that helps identify the different ways a large group of people thinks about a divisive or complicated topic.
- Participants: The people who participated in the conversation by voting and writing statements. Based on how they voted, each participant is sorted into an opinion group.
- Statements: Participants may submit statements for other participants to vote on. Statements are assigned a number in the order they’re submitted.
- Opinion groups: Groups are made of participants who voted similarly to each other, and differently from the other groups.
Talk to the City
"Talk to the City" uses cluster analysis and data visualization for real-time feedback on discussions, spotlighting key opinions. It helps participants spot opinion groups and stick to main topics. Afterward, AI text tools create transcripts, document input, and quickly summarize conclusions. This AI integration highlights tech-driven deliberative workshops as the future of democracy.
Outcomes and Policy Influence
Three Pillars of AI Democratization: Impacts and Strategic Responses
AI is set to fundamentally reshape workflows and job landscapes, raising critical questions regarding the fair distribution of the "labor liberation dividend." To navigate these professional and lifestyle transitions, individuals must engage in continuous learning to acquire foundational technical skills and information literacy. Strategic responses should focus on enhancing the professional capacities of both individuals and local communities, providing state-led educational and vocational training, and expanding diverse pathways for civic participation to ensure that technology serves as a tool for genuine empowerment.
Stakeholder & Industry Level: Industrial Upgrading and Transition Support
While sectors such as Information and Communications Technology (ICT) may transition seamlessly into an AI-driven era, other fields—including design, translation, content creation, legal services, education, and media production—face significant structural disruption. To address these challenges, it is essential to establish robust digital skill-building mechanisms for workers, co-led by professional communities and the state. Furthermore, fostering Public-Private Partnerships (PPP) and community-led initiatives is crucial to assist the public sector in maintaining effective oversight and ensuring necessary checks and balances during this industrial transition.
National Level : Institutionalization of Regulatory Policy
At the national level, the vision is to establish regulations that provide legal certainty and clarity, ensuring that innovation and R&D can flourish without the burden of ambiguity or legal risk. Key response strategies include shaping institutional frameworks with clear compliance paths for diverse stakeholders and establishing responsive feedback and amendment mechanisms to adapt to rapid technological shifts. By engaging key stakeholders through MODA, CIDS maintain a continuous stream of insights into industry-specific execution, fostering a more responsive policy environment. Furthermore, by conducting workshops tailored to specific industrial sectors, CIDS bridges the gap between technology and policy, fostering a robust linkage between public and private sectors to co-create sustainable AI governance. Ultimately, the scope of discussion transitions from the broad societal impacts of AI toward focused policy instruments—such as public governance, civic engagement, and open-call procedures—enabling participants to develop a foundational understanding of the multifaceted impacts AI brings to Taiwanese society.
Note
There's so much more to explore — click here to continue your journey!
The MODA article on “AI-Driven Democratization Workshop”:
https://moda.gov.tw/major-policies/alignment-assemblies/2023-ideathon/1459#toTop
The CIDS article on bridging AI in governance:
https://cid.nccu.edu.tw/genai-future/
The Opening Speech from Audrey Tang (Former Minister of MODA, 2022-2024):
https://pdis.nat.gov.tw/zh-TW/blog/%E6%95%B8%E4%BD%8D%E6%B0%91%E4%B8%BB-AI-%E6%99%82%E4%BB%A3/
Polis Event site: https://polis.tw/2bahudkd2j
Intro. of Talk to the City (T3C): https://www.talktothe.city/
