As local news organizations across the country lose reporters and reduce local government coverage, some organizations are exploring technology and artificial intelligence (AI) to help make public information more accessible.

Local governments make decisions that affect residents every day, from taxes and road projects to school policies and public safety. Keeping up with those decisions often means sitting through hours-long public meetings and sorting through hundreds of pages of agendas and supporting documents.

At the same time, fewer journalists are available to attend those meetings across the country. As AI becomes more common in newsrooms, some organizations are developing tools to decipher public information and make it more accessible.

One of those efforts is LocalMatters, a website that uses AI to transcribe, organize and summarize local government meetings. CU-CitizenAccess.org, which utilizes AI tools and techniques for research and data analysis for stories, is working closely with LocalMatters on developing the best possible resources for journalists.

The platform currently covers public city meetings in Champaign and Urbana and includes a private newsroom product that covers Rantoul and Danville as well. It is also being tested in Fort Wayne, Ind.

Rather than replacing official records, the platform analyzes meeting transcripts, agendas and supporting documents to help users identify topics, follow discussions and navigate primary source material more efficiently.

LocalMatters was founded by impact investor Ben Kramer and software engineer Blair Beverly. 

Kramer said the idea grew from a simple observation: while national politics often dominate public attention, the decisions that shape residents’ daily lives are made much closer to home.

“Your day-to-day life is really centered around local government, but it’s super hard to get involved,” Kramer said. “The breadth of what’s happening locally is really hard to access in an understandable way.” 

Kramer said even residents who want to follow local government can face practical barriers. 

“The meetings last three or four hours,” Kramer said. “There’s agendas that are pretty technical. The supporting documents and packets that come with every meeting can be like hundreds of pages long. It’s definitely a lot of information, way too much, I think, for just someone who’s trying to dabble and get a sense of what’s going on.”

Rather than simply summarizing meetings, Kramer said LocalMatters is designed to identify the issues being discussed and provide readers with enough context to understand why they matter.

“There are other solutions out there that summarize meetings, but it’s often like it’s literally a summary of what is happening and it’s not really helpful to someone if you say, well, docket three, four, and five is approved,” Kramer said. “It doesn’t help me because I don’t know what that docket is or what it really means and why it matters.”

Beverly said the platform draws from the primary source documents that accompany public meetings.

“We’re trying to be more issue-focused,” Beverly said. “The only way you can get the full context is if you’re reading the accompanying documents and packets, which again can be hundreds of pages, and so, kind of integrating that and making it a little easier to navigate.”

Beverly, who previously worked on AI projects at Google, said he has spent the past several months building an automated system that regularly retrieves public meeting videos and documents. This system is what transcribes meetings and associates transcripts with agendas and supporting materials before AI analyzes the content.

Beverly said the AI behind LocalMatters is designed to run one time and work from those source materials rather than generate information on its own or through user prompts.

“One of the most important things is grounding whatever you’re asking of AI in a specific document, a specific transcript, whatever it is, to ensure that it’s not trying to invent a lot of things that it could find in its training data,” Beverly said. “We’re not really asking it to talk about its thoughts or opinions around an issue. It’s more just like ‘what happened in this transcript.'”

A 2025 report by the Reuters Institute for the Study of Journalism and the University of Oxford found that people were considerably more comfortable with journalists using AI as a reporting tool than with AI producing news on its own. 

Across the six countries surveyed, 43% of respondents said they were comfortable when a human journalist led the reporting while using AI as a tool, compared with 21% when AI assisted but a human remained involved and 12% when news was produced entirely by AI. By comparison, 62% said they were comfortable with news produced entirely by human journalists.

The same report found respondents were generally more accepting of AI for tasks such as transcription, translation, editing, headline suggestions and data analysis than for writing complete news stories.

Beverly said LocalMatters is not intended to replace journalists or traditional local news outlets. Instead, he views the platform as a tool that can help both reporters and residents navigate large amounts of public information more efficiently.

“So I do feel like it’s really important that journalists are able to continue to be the people who are interpreting and following up and building the full context for a lot of these stories,” Beverly said. “We are trying to make something that can be helpful for journalists to do that because local news is under-resourced. It’s not possible for a journalist to sit at every one of these meetings and take notes and then follow up. There just isn’t the headcount for it.”

He said the newsroom version of LocalMatters is designed to help journalists identify topics, locate relevant sections of meetings and quickly find quotations from public meetings.

Still, Beverly draws a distinction between what the platform produces and journalism itself.

“I don’t know if I would characterize the output that we have as journalism exactly,” Beverly said. “I think we can be a good tool for journalists who are then going to have their own layer that they’re going to add on.”

“Everything that you see in LocalMatters is basically from the meeting. It’s either from the packet or from the transcript that comes from the meeting. There’s very little extra analysis or interpretation … Journalism to me is bringing the full context and knowing if something is confusing about what’s happening in the meeting. At the moment, we don’t have a way to help you with that. That’s what a journalist would do.”