What governments should know about digital participation & AI
/
Co-authored by Melissa Zisengwe, People Powered and Matt Stempeck, Civic Tech Field Guide
What do a woman from a conservative community in Kyrgyzstan and a UK resident interested in shaping the future of healthcare have in common? They both used digital participation platforms to weigh in on local issues and shape the future of their communities. Digital participation platforms are a vital component of participatory democracy, enabling governments and organizations to engage people in collective activities such as deliberation, consultation, public planning, budget voting, and more.
Digital participation tools are democratic innovations that have emerged over the last two decades to help governments and organizations involve the public in decision-making and addressing public issues in a legitimate, effective, and efficient manner. The emergence of artificial intelligence (AI) and AI-enabled digital participation platforms brings both opportunities and challenges for participatory and deliberative programs.
As digital participation and AI gain importance in participatory processes, governments and organizations require guidance to navigate this space, particularly in selecting and utilizing the appropriate digital tools for their specific contexts. In February, we released our latest digital participation ratings, in which we evaluated 30 digital participation tools (some of which are AI-enabled digital participation platforms) and profiled an additional 17 AI-powered tools.
In this post, we share insights and trends from our research and ratings, including how platforms integrate AI and what governments should consider when approaching this new technology.
The world of digital participation platforms
The top-rated tools in this edition are mostly open-source, including Your Priorities, CONSUL, Decidim, adhocracy+, Democracia OS, LiquidFeedback, and Pol.is, as well as some proprietary platforms, such as Go Vocal, Loomio, and Cap Collectif. These platforms score highly across six key evaluation criteria. Experts on our digital participation committee find them to be cost-effective, reliable, accessible, and adaptable, with minimal technical requirements, comprehensive features, robust data protection, and transparent moderation.
Digital tools for participatory democracy and citizen engagement are now used in most countries, including by all levels of government, universities, international organizations, nongovernmental organizations, foundations, corporations, unions, and other political institutions. These tools have been used for climate action, public space co-design, municipal budgets, gender issues, healthcare, housing, infrastructure, education, waste management, policymaking, and more.
According to our research, most platforms are headquartered in the United States, Canada, and Europe, particularly the United Kingdom, and they have been mostly used in Europe, North America, Oceania, and South America. Several have expanded to Africa and Asia, such as Your Priorities (Thailand), Consul (Somalia), Decidim (Japan), Go Vocal (South Africa), Decision21 (Zambia), and Pol.is (Pakistan). Some remain focused on specific countries, such as Cocoriko, Courbanize, and Open Stad.
Pricing ranges from free for open-source platforms to $20,000 per year for proprietary tools, though some platforms do not disclose costs publicly, requiring potential customers to request more information (and often, a sales call).
Most platforms were launched between 2008 and 2019, with Delib (founded in 2001) being the longest-running in our sample. EngagementHQ, Your Priorities, LiquidFeedback, Ethelo, and Consider.it were all established between 2007 and 2011. English is the most frequently supported language, followed by French, Spanish, and German, though most platforms support additional translations.
Most of the platforms support participatory processes that can be "method-focused" (e.g., deliberation, participatory budgeting) and/or "goal-focused" (e.g., consultation, planning, consensus-building). Of the 30 rated platforms, the majority are designed for deliberative processes, followed by participatory budgeting, and about half support public consultations, participatory policymaking, and planning.
AI meets digital participation: Highlights and trends
Over time, AI technologies have been integrated into most of the platforms we rated. Although many platforms previously utilized AI and Natural Language Processing, with the increased popularity of generative AI, they have begun openly leveraging the technology to introduce new features that enhance participatory processes. Consequently, governments and organizations are increasingly curious about the promises and challenges of AI in democratic processes. Based on our research into 30 platforms that use AI, here’s what governments and organizations should know.
Makeup of AI tools
Digital participation platform developers are actively building, testing, and launching significant AI features and functionalities. Some platforms, such as Crowdsmart, center AI at their core, while others offer AI as an add-on or integrate existing mainstream services like ChatGPT. Notable platforms, including Adhocracy+, Make.org, Empurrando Juntas, Decidim, Consul, 76 Engage, Social PinPoint, EngagementHQ, and Go Vocal, are all engaged in this work. However, in many cases, the generative AI features are still in early development and remain experimental.
Slides from our recent AI and Digital Democracy webinar on opportunities and risks
Some platforms report that while they are innovating and exploring ways to utilize AI, they are concerned about the environmental impact of AI systems, as the energy costs required to power AI can compromise sustainability goals. Developers are also exploring issues of data privacy, ethical considerations, and unintended biases, since the effectiveness of AI models heavily depends on the quantity and quality of the data they are trained on. At the same time, participatory platform developers recognize AI’s potential to enhance efficiency, facilitate data-driven decision-making, increase accessibility, and promote inclusiveness in participatory processes. For example, AI is already helping reduce language barriers in participatory programs.
Most platforms emphasize keeping "humans in the loop" with their AI features and allow program administrators to opt out of using AI. Human oversight helps ensure inclusive results and limits the effects of algorithmic biases. In most cases, AI appears to be primarily used for post-participation analysis rather than powering core participatory activities, with a few exceptions. When AI supports participation, it typically appears as composition box assistance, suggestions, or a supportive chatbot to aid participants. Natural Language Processing-based AI features, such as topic extraction, sentiment analysis, and classification, are common and predate the generative AI boom.
Supported features and functionalities
The majority of the features assist participants and/or administrators by interacting with participants' text contributions in some way, such as translation, summary, classification, synthesis, and moderation. While AI features are often designed to support or assist program administrators (e.g., with overview dashboards or summaries of activity), some features, such as chatbots or “AI user assistants,” are designed to help participants. For example, a chat agent might help guide citizens through participatory processes and assist them if they get stuck at any point in the process.
AI translation is the most commonly used AI feature, appearing in 19 out of the 30 platforms we researched. This demonstrates the application's appeal for participation platforms to reach as many communities as possible. It is commonly used to translate materials, such as learning documents, invitation letters, and social media posts, which can then be verified and improved upon by humans.
The next most common AI feature is sentiment analysis, which enables admins to understand the overall emotional tone (positive, negative, or neutral sentiment) of participants from large amounts of feedback. While clustering techniques were found in over a third of platforms, group similar ideas and user-submitted content. Other common AI functions include reducing toxic content, moderating discussions, summarizing lengthy texts, and guiding users through participatory processes with AI chatbots or assistants. Some platforms have launched more advanced generative AI features, such as visualizing physical places within their communities through image generation or assisting in the formation of representative groups.
Here are some notable uses of AI in the platforms we reviewed:
All Our Ideas and Assembl (Bluenove) apply generative AI across their entire platforms, helping participants make sense of discussions through narrative tools, mindmaps, and storytelling techniques.
UrbanistAI simulates policy outcomes by visualizing the effects of interventions, such as changes to traffic flow and air quality, on community health.
Make.org integrates a Bayesian learning algorithm into its participatory process to ensure randomized and fair proposal selection for voting.
Polco deploys its AI chatbot, Polly, to analyze participation data, generate strategic insights, and assist with content creation, grant writing, and organizational planning.
PlaceSpeak, deliberAIde, and the Stanford Online Deliberation Platform focus on content quality and moderation, utilizing AI for sentiment analysis, the removal of personal information, user nudging, and maintaining discussions aligned with agendas.
What governments and organizations should consider when using AI-powered digital participation platforms
First, admins should ensure that any AI tool they select is aligned with any AI principles or rules their institution has implemented, both in spirit and in practice.
Next, governments and organizations need to consider their internal capacity requirements to use the tool independently. They should assess whether AI-specific training is provided to administrators and end-users alike, as setting up an AI tool may require a high level of team engagement, making administrators highly dependent on the product owners.
Any application of AI must support meaningful public input and be transparent to the user, i.e., the user must be informed that they are using an AI system. It should also be clear how AI is used in decision-making, the data being used, and what role AI plays in shaping outcomes.
When choosing an AI-powered participation platform, look for those that minimize bias and errors, especially those trained on specialized data.
Prioritize tools that empower citizens to shape decisions rather than replace human participation. The ordinary citizen is typically a novice in using digital participation tools, let alone AI. Therefore, effective participation requires governments and organizations to educate and empower the public on how these tools work.
AI translation is generally effective for high-resource languages – languages with rich data resources available, such as English – but less effective for translating low-resource languages. Given that the vast majority of languages in the world are under-resourced, this discrepancy could exacerbate inequalities within participatory processes, with speakers of high-resource languages receiving higher-quality translations compared to speakers of low-resource languages.
In terms of ethics and transparency, admins need to understand how AI is being used on the platform. This includes algorithmic transparency — can administrators understand how results are generated? These considerations can help you prepare for and address potential biases in the outputs produced by AI systems, as well as accessibility limitations for your community.
Call for case studies: Share your experience with AI-powered digital participation tools!
Following the release of our Digital Participation Tools ratings in February, we are developing detailed case studies to highlight best practices worldwide. We are currently identifying examples of digital participation and AI applications that have resulted in clear benefits for governments, organizations, and communities. Our goal is to research, develop, and share key examples in which digital participation and/or AI tools have been applied to participation and deliberation processes, evaluating both the benefits and the challenges faced along the way.
Help us by sharing your experiences by May 26, 2025.
Looking for more guidance?
We have developed several resources to assist governments and organizations in selecting and utilizing digital tools for public participation.