Google is putting artificial intelligence at the center of a new effort to help governments understand what their communities are saying, supporting technology designed to process large-scale public conversations and make it easier for officials to identify common concerns, areas of disagreement and emerging priorities.
Image Courtesy : ki-company.ai
The initiative builds on work by Jigsaw, Google's technology incubator focused on projects related to information, security and human agency. Jigsaw has been experimenting with AI-powered “sensemaking” tools that can analyze thousands of individual contributions and turn them into more understandable summaries of what a community is discussing.
The effort reflects a broader question emerging as generative AI becomes more capable: Can artificial intelligence help governments listen to residents at a scale that traditional town halls, surveys and focus groups cannot easily accommodate?
Jigsaw believes it can.
The organization describes the problem as a deficit in what it calls “structural agency”—a gap between people's desire to influence the systems and decisions affecting their lives and their ability to do so. Jigsaw says AI could potentially help narrow that gap by making it easier for communities to participate in large conversations and for decision-makers to understand the resulting information.
The technology is not intended simply to create another government chatbot. Instead, Jigsaw's sensemaking approach focuses on analyzing conversations that have already taken place.
That distinction is important.
A conventional public meeting might attract dozens or hundreds of participants, while an online consultation can potentially generate thousands of individual comments and ideas. The larger the conversation becomes, however, the harder it can be for government officials to manually review every contribution and determine what themes are emerging.
AI can potentially address that information bottleneck.
Jigsaw has developed an open-source sensemaking library that can categorize large collections of public input, identify themes and summarize the information. The technology has already been tested in Bowling Green, Kentucky, where local leaders used it as part of a long-term community planning initiative.
The Bowling Green project began with a community conversation about the future of Warren County as the region prepared for significant population growth. Nearly 8,000 residents participated, submitting thousands of ideas that received more than one million responses from other community members. Jigsaw's technology was then used to help organize and summarize the resulting material.
The scale of that experiment illustrates why AI could be attractive to local governments.
Manually reading and categorizing thousands of comments can require significant amounts of staff time. According to a Jigsaw case study, 90% of surveyed local leaders involved in the Bowling Green project said the resulting analysis gave them a better understanding of what the community wanted. The leaders also estimated that the technology saved an average of 28 days of manual work.
For government agencies, that could represent a meaningful operational change.
Instead of treating public engagement as a meeting that happens at a particular time and location, officials could potentially create continuous digital conversations in which residents submit ideas, respond to other residents and express areas of agreement or disagreement.
AI could then help transform that enormous collection of information into something government staff can actually examine.
Jigsaw has been developing this concept for several years.
In 2024, the organization described a broader effort to use large language models to expand deliberative technology. The goal was to combine the breadth of large-scale polling with some of the depth normally associated with focus groups, allowing people to express their views in their own words while giving communities and decision-makers tools to understand the resulting discussion.
That work eventually expanded into additional experiments.
Jigsaw's “We the People” initiative, for example, involved conversations with more than 2,500 Americans from across the country. Participants were able to discuss issues, respond to ideas expressed by others and vote on propositions generated from the broader conversation. Jigsaw said more than 90% of participants felt their opinions were represented in the resulting summary, while nearly 80% said the process helped them better understand other people's perspectives.
The underlying technology is therefore being positioned as a form of AI-assisted public listening rather than AI-based government decision-making.
That distinction could become increasingly important as governments experiment with artificial intelligence.
There is a fundamental difference between using AI to organize public input and allowing an AI system to determine government policy. In the first model, the technology acts as an analytical tool that helps humans process information. Officials and residents remain responsible for deciding what actions should follow.
Jigsaw's public description of the technology emphasizes this role.
The organization says its goal is to help people have greater agency and make community voices easier to understand. Its sensemaking work includes capabilities such as summarization, visualization, topic analysis and identifying areas where opinions converge or diverge.
Google.org has separately expanded its investment in AI for government.
In September 2026, Google announced 15 recipients of its $30 million Google.org Impact Challenge: AI for Government Innovation. The selected organizations include academic institutions, social enterprises and nonprofits developing AI applications for public-sector challenges. Google said the recipients would receive funding as well as technical assistance from Google engineers and product specialists through a dedicated accelerator.
The program illustrates how Google's philanthropic and technology organizations are approaching government AI from multiple directions.
Some projects focus on improving government services, while Jigsaw's work focuses more directly on information flowing between communities and decision-makers.
Google Cloud has also been promoting AI applications for state and local government operations, including tools designed to reduce administrative workloads, connect fragmented information and automate routine processes. Google Public Sector said in September that AI is increasingly becoming a central priority for government technology leaders as agencies attempt to modernize legacy systems and improve service delivery.
Together, these efforts point toward a broader expansion of AI throughout government.
But Jigsaw's approach tackles a different problem from simply automating paperwork.
It is attempting to improve the government's ability to understand large amounts of human feedback.
That becomes particularly relevant as cities and counties confront complicated issues that can generate thousands of opinions. Housing, transportation, infrastructure, development, environmental planning and other local decisions can produce extensive public debate.
Traditional engagement methods often struggle to capture the full range of those perspectives.
A public meeting might disproportionately attract residents who have the time and ability to attend. A survey can reach more people but generally requires researchers to decide the questions in advance. A conventional focus group can generate detailed qualitative information but is typically limited to a relatively small number of participants.
AI-assisted deliberation could combine elements of each approach.
Residents can potentially speak in their own words, interact with other participants and raise ideas that officials did not anticipate when designing a survey.
The AI then has the job of helping humans make sense of the resulting discussion.
That could change the economics of public engagement.
If a local government can process 10,000 or 50,000 individual contributions without requiring a corresponding increase in staff devoted solely to reading and categorizing them, larger-scale public consultations could become more practical.
However, the technology also introduces new questions.
AI-generated summaries are interpretations of underlying information. Even when a system is designed to identify themes rather than make policy recommendations, the process of deciding which themes are important can affect how the conversation is presented.
A summary can potentially omit a minority viewpoint, combine distinct opinions into a single category or give greater visibility to frequently mentioned ideas than to less common but significant concerns.
That means transparency becomes an important part of AI-assisted public engagement.
Residents may reasonably want to know how their contributions were analyzed, what information was included, what was excluded and how AI-generated conclusions were reviewed by humans.
Jigsaw's emphasis on open-source technology could be significant in this context.
The organization's sensemaking library has been made available as open-source technology, allowing others to examine and build on the underlying tools rather than relying exclusively on a proprietary black-box system. Jigsaw says it intends to continue sharing research, insights, datasets and code as its work develops.
Open technology does not eliminate every concern surrounding AI-generated analysis, but it can provide additional opportunities for independent examination and adaptation.
There is also the question of privacy.
Large-scale public conversations can contain personal information, sensitive experiences and other material that residents may not expect to be processed by an AI system. Any government or civic organization deploying such technology therefore has to consider what information is collected, how it is stored and who can access it.
These considerations become even more important as AI systems become capable of extracting relationships and patterns from enormous amounts of text.
Jigsaw's work is taking place against a backdrop in which governments around the world are already attempting to determine how AI should be incorporated into public institutions.
At the same time, Google is promoting AI tools for government through its broader Google.org and Google Public Sector initiatives. The company's $30 million AI for Government Innovation challenge demonstrates the scale of that investment, with 15 organizations selected from more than 2,600 proposals submitted globally.
The potential payoff is substantial.
Governments possess enormous amounts of information, but information alone does not guarantee that officials can understand what communities want. A city might receive thousands of comments about a proposed development, for example, but those comments can be difficult to categorize and compare.
An AI system capable of rapidly identifying themes could provide a different way of looking at the information.
Instead of simply counting how many people support or oppose an issue, the technology could potentially reveal why people hold particular views, which ideas receive support across otherwise different groups and where disagreements are concentrated.
That is one of the reasons Jigsaw describes its approach as “sensemaking.”
The objective is not simply to count voices.
It is to understand them.
That could be particularly useful in situations where a community's views are more complicated than a simple yes-or-no question.
A transportation project, for example, might receive broad support while residents disagree about its route. A housing plan could have support for additional housing but disagreement about building height. A redevelopment proposal might receive support for economic investment while generating concerns about affordability.
Traditional polling can struggle to capture those nuances.
A large-scale deliberative conversation can produce them, but only if someone has a practical way to process the information.
AI potentially provides that processing layer.
Jigsaw's experiments suggest that this could allow governments to engage communities at a much larger scale than conventional public meetings.
The technology could also potentially make public engagement more continuous.
Rather than asking residents to attend a single town hall, governments could maintain digital conversations over longer periods. Residents could see what others are saying, respond to proposals and refine their own views as they encounter different perspectives.
That creates the possibility of a more dynamic feedback loop between citizens and institutions.
But it also means governments will need to establish clear rules around how AI-generated insights are used.
An AI summary should not automatically become a substitute for democratic processes, elected representatives, professional expertise or legally required public procedures.
Instead, it could become another source of information available to the people responsible for making decisions.
That distinction may ultimately determine whether AI-powered civic engagement becomes a lasting government technology or simply another short-lived experiment.
For Google and Jigsaw, the opportunity is broader than building another AI application.
It is an attempt to use increasingly capable models to address a longstanding problem: how institutions can listen to large populations without reducing complex human opinions to a handful of survey questions or the loudest voices in a room.
The Bowling Green experiment provides an early example of what that might look like in practice. Thousands of residents contributed ideas, those ideas generated more than a million interactions, and AI-assisted analysis helped local leaders turn the resulting information into a more manageable picture of the community's views.
The next phase will determine whether that approach can work across different communities, political environments and government structures.
If it can, AI-assisted sensemaking could become a new layer of civic infrastructure—one that sits between large-scale public conversations and the humans responsible for interpreting them.
Google is already expanding its investments in AI for government, while Jigsaw continues experimenting with technologies intended to increase public participation and make large conversations easier to understand.
The larger experiment is therefore underway.
Artificial intelligence has spent the past several years becoming better at generating information. Google's civic AI efforts are exploring a different possibility: using AI to help institutions understand information coming from people.
Whether that ultimately produces more meaningful public participation will depend not only on the capabilities of the technology, but also on how governments, communities and residents choose to use it.
For now, Jigsaw's work suggests that one of AI's emerging roles may be less about speaking for people and more about helping institutions hear what large numbers of people are already saying.