POLIS: The AI Platform Shaping Democratic Decision-Making Worldwide

From its origins in the United States to its widespread adoption in Taiwan, the United Kingdom, Finland and beyond, POLIS is emerging as a powerful tool for fostering consensus and informing policy decisions. This open-source artificial intelligence platform aggregates opinions from large populations, identifying common ground and bridging divides in a way traditional methods often fail to achieve.

Origins and Development of POLIS

POLIS was initially developed in the USA, with its core software still accessible on the original POLIS website. In 2018, the Computational Democracy Project was founded as a non-profit organization to continue the open-source development and expand the platform’s reach. The project is committed to designing fairer, decentralized democracies through advanced statistics and machine learning. POLIS 2.0, the latest version, leverages distributed cloud infrastructure, real-time summarization, automatic translation using Large Language Models (LLM), and advanced topic analysis through the EVōC Library.

How POLIS Works: Beyond the Reply Button

POLIS distinguishes itself from conventional online discussion forums by intentionally eliminating the “reply button.” This design choice structurally prevents flame wars and unproductive arguments. The platform operates on a “voting matrix” where participants respond to short opinions with “agree,” “disagree,” or “pass.” Machine learning and statistical algorithms then analyze this data, identifying key themes and automatically generating groups based on shared viewpoints. The system utilizes Principal Component Analysis (PCA) and k-means methods to visualize opinion distribution.

Taiwan: A Pioneering Case Study

Taiwan first implemented POLIS in March 2014, spurred by the sunflower student movement and a need for improved citizen engagement. Digital Minister Audrey Tan spearheaded the “vTaiwan” initiative, integrating POLIS as a central tool for large-scale public participation. This led to significant policy outcomes, including regulations for Uber based on a 95% consensus around “passenger safety.” POLIS has since played a crucial role in formulating over a dozen other bills and regulations in Taiwan, covering issues like anti-revenge porn and fintech regulations, establishing it as a core component of the nation’s democratic infrastructure.

Global Applications of POLIS

Beyond Taiwan, POLIS is being utilized in diverse contexts worldwide:

• Finland: Designing safety support for the elderly and mental health services for children based on resident feedback.
• United Kingdom: National security consultation and local government parking policy development.
• Bhutan, Timor-Leste, and Pakistan: The United Nations Development Program (UNDP) deployed POLIS in the largest online deliberation exercise in history, involving 30,000 young people.
• Other Locations: Governments of Singapore and the Philippines, the Citizens Council on Climate in Austria, and the city of Bowling Green in the United States are similarly leveraging POLIS to address complex challenges.

Technical Specifications and Accessibility

POLIS is published under an AGPL-3.0 license and is built using languages including JavaScript, Python, TypeScript, and Clojure. It utilizes Docker infrastructure and incorporates security measures like SSL certificate settings and JWT key generation for OIDC authentication.

Key Takeaways

• POLIS is an open-source AI platform designed to facilitate consensus-building in large groups.
• Its unique design, eliminating the “reply button,” minimizes conflict and promotes constructive dialogue.
• Taiwan has successfully integrated POLIS into its democratic processes, influencing legislation on various issues.
• The platform is being adopted globally for diverse applications, from national security to local policy development.