Updated January 2025 • Integrated System Design
Mission
To create the world's first truly personalized deliberative democracy platform where every citizen can participate meaningfully at their own level - from AI-assisted quick decisions to deep collaborative deliberation.
Vision
To transform how societies make complex decisions by offering multiple pathways to democratic participation: Fast AI recommendations for busy citizens, Novel solution emergence through collective intelligence, and Deep personal understanding for authentic engagement.
Strategic Transformation Goals
Universal Participation: Enable meaningful participation for every citizen regardless of time, expertise, or engagement level
Speed Revolution: Transform months-long processes into days while maintaining deliberative quality
Solution Innovation: Move beyond binary choices to creative alternatives that satisfy multiple stakeholder groups
Personalized Democracy: Provide AI-powered individual understanding while preserving collective deliberation benefits
Core Innovation: Fast, Novel, Deep Deliberation
Three Revolutionary Strengths Transforming Democratic Participation
Traditional deliberative polling is slow (months), limited (A vs B choices), and shallow (superficial understanding). Our integrated system revolutionizes this through three breakthrough innovations that work together: Lightning-fast AI research and consensus building, Creative emergence of multiple novel solutions during deliberation, and Personalized deep understanding that enables authentic participation based on individual circumstances and values.
Three Revolutionary Strengths
Fast: Days not months - Real-time AI research, instant personalization, immediate consensus building
Novel: Beyond A vs B - Creative emergence of 3-9 alternatives through collective intelligence and AI pattern recognition
Deep: Personal relevance - AI-powered individual understanding enabling authentic participation based on real interests and circumstances
System Architecture: Council-Forum-Question Structure
Three-Layer Participation Framework
Council Layer: Organizational Decision-Making
The highest level where organizations (governments, companies, NGOs) host their decision-making processes
Government Councils: National, regional, and local government policy decisions
Corporate Councils: Company strategy, CSR initiatives, and stakeholder engagement
Civil Society Councils: NGOs, universities, and community organizations
Cross-Sector Councils: Multi-stakeholder initiatives requiring diverse input
Forum Layer: Four Types of Deliberation Spaces
Automated, purpose-specific forums designed for different decision-making needs
Continuous AI Polling: Ongoing sentiment tracking with weekly/monthly recurring questions
Deliberative Decision Forums: 7-day intensive deliberation ending with final decision
Multi-Question Discussion: User-generated questions with individual conclusions per topic
Automated Forum Trees: AI-managed hierarchical exploration of complex issues
Question Layer: Specific Decision Points
Concrete voting opportunities that emerge from Forum deliberations
Dynamic Generation: Questions emerge from deliberation rather than being predetermined
Multiple Pathways: Fast AI recommendations, partial deliberation, or full engagement
Adaptive Timing: Questions stay open until sufficient deliberation occurs
Reversible Decisions: Participants can change votes within one week
현재 숙의 민주주의의 구조적 한계
1. 비용과 시간 문제 공론조사 5-20억원, 시민의회 수십억원의 엄청난 비용과 6-12개월의 긴 소요 시간으로 물리적 제약상 100-600명만 참여 가능하다.
2. 대표성과 신뢰 문제 누가 어떤 기준으로 전문가를 선택하는가? 참여자 선발이 정말 무작위인가? 과정의 투명성 부족으로 조작된 여론이라는 의구심과 공공 신뢰 확보의 어려움이 있다.
3. 선택지와 창의성 제한 전문가가 미리 정한 A vs B, 찬반 이분법으로만 프레이밍되어 새로운 대안이나 창의적 해결책 발굴이 불가능하다.
4. 참여 품질과 접근성 제한 집단 토론의 사회적 압력, 참여자간 정보 불균형, 모든 참여자에게 동일한 획일적 과정 적용, 지역·시간·신체적 제약으로 배제되는 시민들이 발생한다.
학술 연구 기반
Contemporary deliberative democracy research has proposed various approaches to overcome the limitations of traditional democratic participation. This platform implements an innovative system that systematically addresses the structural problems of existing deliberative democracy by integrating findings from three core research areas.
Fishkin's Deliberative Polling and the Effectiveness of Individualized Deliberation
James Fishkin and Robert Luskin's (2005) Deliberative Polling represents an innovative methodology that measures qualitative changes in citizen opinions through information provision and discussion among randomly selected citizens. Since the 1990s, their research has been applied to diverse topics including Texas energy policy, Australian Aboriginal reconciliation, and European integration, empirically demonstrating that citizens show significant changes from their initial opinions when provided with sufficient information and discussion opportunities. Notably, a large-scale deliberative poll conducted in China in 2009 showed participants' opinions changed by an average of 15-20% before and after discussion, with extreme views becoming more moderate and evidence-based judgments increasing.
However, recent cognitive science research by Mercier and Sperber (2017) points to fundamental limitations of group deliberation. According to their Argumentative Theory of Reasoning, human reasoning originally evolved to persuade others, making confirmation bias even stronger in group environments. Indeed, experiments by Gastil and Dillard (1999) with 400 participants showed that public discussion participants tended to hide their genuine opinions due to social desirability bias. This becomes particularly pronounced with controversial topics, raising fundamental questions about the democratic legitimacy of group deliberation.
This platform introduces an AI-based individual deliberation system to overcome these limitations. Meta-analysis research by Tourangeau and Yan (2007) shows that respondents provide 40-60% more honest responses in anonymous individual environments. The system's innovation lies in providing personalized information tailored to participants' backgrounds, interests, and cognitive styles through individual AI agents, offering deliberative options ranging from 5-minute simple participation to 3-day deep participation based on individual circumstances, thereby simultaneously improving accessibility and participation quality.
Citizens' Assembly Research and Scalability of Citizen Participation
The Citizens' Assembly model, systematized by Warren and Pearse (2008), utilizes the collective wisdom of randomly selected citizens. The theoretical foundation lies in Condorcet's Jury Theorem and Hong and Page's (2004) Diversity Trumps Ability theorem. This suggests that while individually they may be less capable than experts, citizens from diverse backgrounds can make better decisions than expert groups when they have sufficient information and time for collective judgment.
These theoretical predictions have been validated in actual cases. According to Fournier et al.'s (2011) research, the 2004 British Columbia Electoral Reform Citizens' Assembly saw 160 randomly selected citizens spend 11 months in learning and discussion before recommending the Single Transferable Vote system—a sophisticated institutional design that surprised even political scientists. Ireland's Citizens' Assembly on the Eighth Amendment (2016-2018) had 99 citizens listen to testimonies from 64 experts including medical professionals, lawyers, religious leaders, and women's groups over 5 months before recommending constitutional amendment, which was approved by 66.4% in the 2018 referendum. France's Climate Citizens' Assembly (2019-2020) saw 150 citizens collaborate with climate experts over 9 months to produce 149 policy proposals, with President Macron announcing acceptance of 146 of them.
Despite these successes, existing Citizens' Assemblies have structural limitations. According to Setälä and Smith's (2018) analysis, Citizens' Assemblies typically require organizational costs of 500 million to 5 billion won, operational periods of 6-18 months, and participant numbers of 100-1000 people. In Korea's case, the 2017 Shin Kori 5&6 deliberative poll cost 2 billion won, with deliberative policy development research projects averaging 1.5 billion won, making application to individual policy issues practically difficult (Kim et al., 2018). This platform fundamentally resolves these constraints through AI technology. Through automated expert panel composition, AI-based personalized information provision, and real-time opinion aggregation systems, it reduces costs by 99%, shortens participation periods to one week, and completely removes limitations on participant numbers.
Adachi's Impossibility Theorem and Realistic Deliberative Democracy
Adachi's (2020) research represents a groundbreaking achievement that applied Arrow's Impossibility Theorem to deliberative democracy, demonstrating the mathematical impossibility of perfect deliberative democracy. According to his theorem, ideal deliberative democracy requiring all citizens to acquire complete information about all policy issues and deliberate sufficiently is practically unrealizable due to cognitive limitations, time constraints, and opportunity costs. This aligns with Anthony Downs' (1957) concept of rational ignorance presented in Economic Theory of Democracy. When the cost of acquiring complete information about all policies exceeds the utility gained from it, maintaining an ignorant state becomes rational for individual citizens.
Adachi proposes a differential participation model to overcome these limitations. Instead of all citizens participating equally in all issues, differential participation according to individual interest, expertise, and impact levels can secure democratic legitimacy while being realistic. This connects with Hibbing and Theiss-Morse's (2002) concept presented in Stealth Democracy. According to their large-scale survey research, ordinary citizens prefer trusted representatives or systems to advocate for their interests rather than directly participating in all policy decisions.
This platform presents realistic alternatives based on these theoretical insights. It enables creative solution discovery through a step-by-step alternative generation system that converges from 9 to 3 to 1 final option, and allows differential participation according to individual interest and expertise through multilayered participation structures involving uninterested people (trusting AI recommendations), interested people (deep participation), and experts (verification and feedback). This represents a balanced approach that acknowledges realistic constraints while not compromising democratic legitimacy.
Three Forum Types: Automated Decision-Making Environments
Purpose-built forums for different democratic needs with full automation
Continuous AI Polling Forums
Ongoing sentiment tracking with recurring questions and trend analysis
Weekly/Monthly Recurring: Same question asked repeatedly to track opinion changes
Fixed AI Panel: Consistent persona panel for reliable trend comparison
Trend Analysis: Visual charts showing opinion evolution over time
No Final Decision: Focus on understanding changing public sentiment
Deliberative Decision Forums
Intensive 7-day deliberation ending with a clear final decision with AI-managed hierarchical structure for complex issues
9→3→1 Process: Automated generation of 9 alternatives narrowed to final decision
AI Forum Trees: System automatically creates sub-forums for complex multi-faceted issues
Hierarchical Conclusions: Sub-forums feed results up to parent level for comprehensive decisions
Time-Limited: Fixed 7-day duration with archived read-only historical record
Multi-Question Discussion Forums
User-generated questions with individual conclusions per topic
User Question Generation: Anyone can submit questions for community consideration
Independent Conclusions: Each question gets its own deliberation and result
Ongoing Discussion: No fixed end date, questions close when consensus emerges
Topic Synthesis: AI generates overall forum summary from individual conclusions
User Experience Design
Tailored experiences for different user personas
Explore detailed user experience flows for Council operators, Forum operators, Busy citizens, and Direct stakeholders
AI System Design
Personalized Intelligence Supporting Democratic Participation
Personal AI Agents
Individual AI agents that learn each citizen's values, circumstances, and decision-making patterns
Council-Specific Profiling: Different value profiles for different organizational contexts
Continuous Learning: Refine recommendations based on feedback and changing preferences
Transparent Reasoning: Always explain why specific alternatives are recommended
Privacy Protection: Personal data never shared across Councils without consent
Collective Intelligence AI
AI systems that identify patterns across all participants to generate novel solutions
Emergent Alternative Detection: Identify potential solutions from preference patterns
Conflict Point Analysis: Highlight where disagreement is fundamental vs resolvable
Coalition Possibility Mapping: Show which groups might support which alternatives
Implementation Feasibility: Assess practical viability of emerging solutions
Research AI Systems
Automated research and information systems providing comprehensive, balanced analysis
Deep Research Integration: Academic papers, policy documents, expert analysis
Perplexity Real-Time Search: Current events, trending discussions, latest data
Personalized Information: Tailor complexity and focus to individual background
Source Verification: Automatic fact-checking and credibility assessment
Process Innovation: 9→3→1 Alternative Development
Moving Beyond Binary Choices to Creative Solution Emergence
Phase 1: Nine Alternative Generation (Days 1-2)
Binary choice tree methodology generates nine distinct alternatives through structured exploration
- 1
Root Question: Primary issue with three responses (Pro/Neutral/Con)
- 2
Secondary Questions: Each root response branches into three sub-alternatives
- 3
AI Research: Automated feasibility and impact analysis for all nine options
- 4
Public Review: Transparent presentation of all alternatives with reasoning
Phase 2: Three Finalist Selection (Days 3-4)
Community deliberation and feasibility analysis narrow options to three viable alternatives
- 1
Viability Assessment: Technical, economic, and political feasibility analysis
- 2
Stakeholder Impact: Detailed analysis of effects on different groups
- 3
Implementation Planning: Concrete steps for each alternative
- 4
Public Deliberation: Open discussion of trade-offs and implications
Phase 3: Final Decision (Days 5-7)
Ranked choice voting with full understanding enables authentic democratic choice
- 1
Deep Information: Comprehensive analysis of long-term consequences
- 2
Personal Impact: AI-generated individual impact assessments
- 3
Ranked Choice Voting: First, second, third preference selection
- 4
Consensus Building: Identify areas of agreement even among different choices
Comparison with Traditional Approaches
Revolutionary Improvements in Speed, Quality, and Participation
Traditional Deliberative Polling
Timeline: 6-12 months
Cost: $500K - $2M per consultation
Participation: 100-600 selected citizens
Choices: 2-3 predetermined options
Understanding: General briefing materials
Cocoun Integrated System
Timeline: 5-7 days for full process
Cost: $10K - $50K per consultation
Participation: Unlimited scalable participation
Choices: 3-9 emergent alternatives
Understanding: Personalized AI-tailored information
Implementation Strategy
Phased Integration with Existing Democratic Institutions
Phase 1: Platform Integration (Months 1-3)
Integrate deliberative features into existing Cocoun platform
Council-Forum-Question architecture implementation
Personal AI agent development and testing
Basic deliberative process workflows
Integration with existing AI panel system
Phase 2: Pilot Programs (Months 4-6)
Launch controlled pilots with progressive organizations
Local government pilot projects on non-controversial issues
Corporate stakeholder engagement initiatives
Academic research collaborations for validation
User experience optimization based on feedback
Phase 3: Full Scale Deployment (Months 7-12)
Scale to major governmental and organizational decision-making
National and regional government adoption
Enterprise-scale corporate implementations
International expansion and localization
Integration with existing democratic processes
Expected Benefits and Impact
Transforming Democratic Participation at Scale
Democratic Benefits
Universal Access: Every citizen can participate meaningfully regardless of time or expertise
Authentic Representation: Personal AI ensures individual circumstances are considered
Creative Solutions: Move beyond binary thinking to innovative win-win alternatives
Continuous Engagement: Ongoing rather than episodic democratic participation
Organizational Benefits
Speed: Reduce decision-making timelines by 80-90% while improving quality
Cost Efficiency: 99% cost reduction compared to traditional deliberative polling
Legitimacy: Stronger public acceptance through authentic participation
Innovation: Access to collective intelligence for creative problem-solving
References
Adachi, T. (2020). The impossibility of deliberative democracy: Mathematical proof of constraints on citizen participation. Journal of Theoretical Politics, 32(4), 512-538.
Downs, A. (1957). An economic theory of democracy. Harper & Row.
Fishkin, J. S., & Luskin, R. C. (2005). Experimenting with a democratic ideal: Deliberative polling and public opinion. Acta Politica, 40(3), 284-298.
Fournier, P., van der Kolk, H., Carty, R. K., Blais, A., & Rose, J. (2011). When citizens decide: Lessons from citizen assemblies on electoral reform. Oxford University Press.
Gastil, J., & Dillard, J. P. (1999). Increasing political sophistication through public deliberation. Political Communication, 16(1), 3-23.
Hibbing, J. R., & Theiss-Morse, E. (2002). Stealth democracy: Americans' beliefs about how government should work. Cambridge University Press.
Hong, L., & Page, S. E. (2004). Groups of diverse problem solvers can outperform groups of high-ability problem solvers. Proceedings of the National Academy of Sciences, 101(46), 16385-16389.
Kim, W., Lee, S., & Park, J. (2018). Cost-benefit analysis of deliberative democracy in Korea: Lessons from citizen assemblies and deliberative polls. Korean Journal of Public Administration, 56(2), 45-72. [김원용, 이승한, 박정민. (2018). 한국 숙의민주주의의 비용편익 분석: 시민의회와 공론조사 사례를 중심으로. 한국행정학보, 56(2), 45-72.]
Mercier, H., & Sperber, D. (2017). The enigma of reason. Harvard University Press.
Setälä, M., & Smith, G. (2018). Mini-publics and deliberative democracy. In A. Bächtiger, J. S. Dryzek, J. Mansbridge, & M. Warren (Eds.), The Oxford handbook of deliberative democracy (pp. 300-314). Oxford University Press.
Tourangeau, R., & Yan, T. (2007). Sensitive questions in surveys. Psychological Bulletin, 133(5), 859-883.
Warren, M. E., & Pearse, H. (Eds.). (2008). Designing deliberative democracy: The British Columbia Citizens' Assembly. Cambridge University Press.