The winning team developed a functioning, scalable prototype for an innovative news platform powered by generative AI
A team of three University of Minnesota computer science students were awarded the $10,000 grand prize at the Minnesota Journalism Center’s first-ever hackathon. The competition brought Minnesota data science and journalism communities together to harness cutting-edge technologies for making local civic information more accessible to more Minnesotans.
Over the span of less than 24 hours, the winning team of coders — master’s students Dipan Bag, Erina Karati, and Arunachalam Manikandan — developed a functioning, scalable prototype for an innovative news platform powered by generative artificial intelligence that they called MinneDigest. It offers locally tailored summaries of news stories published by Minnesota news outlets which can be repackaged into instant audio formats or translated into multiple languages to reach non-English-speaking audiences. The team said they hoped that MinneDigest could make it easier for more Minnesotans to discover and engage with local journalism.
The Journalism Center hosted the hackathon — in partnership with the nonprofit Hacks/Hackers organization — at Murphy Hall, home to the Hubbard School of Journalism & Mass Communication. The event drew nearly 80 participants from across the university and the professional journalism and tech communities in the Twin Cities and beyond. The event was part of a multi-day series of training, workshops and conversations the Minnesota Journalism Center hosted in late January about AI and journalism — on campus and at the Minnesota Newspaper Association's annual convention in Brooklyn Park, Minn.
“I know this place is already home to vibrant communities in local journalism and data science, to say nothing of our incredibly engaged civic groups. In holding this event, we hoped these groups would not only come together and connect but work alongside each other around a common mission, and I was truly inspired by the enthusiasm of those who participated.” —Benjamin Toff, director of the Minnesota Journalism Center
Generating ideas for generative AI at the MNA Convention
The winning team said they drew partial inspiration from one of four 60-second videos that hackathon organizers played during the event's kickoff, in an effort to encourage participants to think about the types of problems and opportunities that journalists and other providers of local information are facing. The videos featured elevator pitches recorded during the previous day's idea-development workshop at the MNA Convention. During the workshop, publishers, editors and other journalists from a broad range of local news outlets generated practical ideas for applying AI to solve their newsrooms' challenges.
In the video that inspired the winning team, Trevor Slette, publisher of Citizen Publishing Company in the southwest Minnesota town of Windom, asked hackathon teams if they might be able to develop an AI tool that would allow him to make his newspaper’s content available to non-English-speaking residents across the region.
Another team said they were inspired by the short video created by Bill Abrecht, regional president of Adams Publishing Group's presence in east-central Minnesota. Albrecht spoke about the difficulties his papers faced in tracking down, confirming and aggregating high school and community sports scores in a timely fashion. In response, a team comprised of data scientist Nick Capaldini, MPR News' Michael Olson, and University of Minnesota social psychology graduate student Anshu Patel and mass communication graduate student Ningyuan Ma, created VeriScore, an AI-powered solution designed to streamline that process.
Other pitches from the idea-generation workshop called for an AI-driven tool that would track and surface relevant local information and impacts embedded in bills and testimony at the Minnesota Legislature, to allow for stronger coverage of the lawmaking process. Another pitch proposed an AI-powered insight tool that small, local newspapers could use to point advertising sales teams toward more informed decisions about possible local leads and business development.
The MNA idea-development workshop featured lightning talks from local journalists who are already incorporating AI into their workflows: The Minnesota Star Tribune's Chase Davis, who leads the newspaper's strategy and transformation efforts, talked about some of the new initiatives he is leading with his team. The paper recently received a grant from the nonprofit Lenfest Institute and OpenAI to experiment with AI summarization, analysis and content discovery. The Sahan Journal's Cynthia Tu, who specializes in AI and data reporting, demonstrated ways she has used AI to conduct innovative investigative projects and enhance her newsroom’s day-to-day operations. Tu and the Sahan team have also been recognized for their AI innovation, through an ongoing partnership with the American Journalism Project.
Lynn Walsh, Assistant Director at the nonprofit training organization Trusting News, shared how local news organizations she is working with around the country are tackling questions of disclosure and labeling, and incorporating AI into their workflows analyzing public data, spotting patterns in community feedback and repackaging content for a variety of distribution platforms.
Centering ethics, trust, equity and audience
On the final day of the MNA convention, MJC director Benjamin Toff, Walsh and Alex Mahadevan, director of the Poynter Institute's MediaWise news literacy project, led a separate conversation focusing on questions of audience trust, disclosure and transparency — and the underlying questions of ethics and equity — in newsrooms' use of AI.
Toff shared findings from his research, including focus groups he held in 2024 with audiences in Minnesota about how they are thinking about the use of AI in news. The findings highlighted the degree to which many feel a combination of anxiety and annoyance around its increasing presence in widening aspects of people’s digital experiences.
In light of this baseline skepticism, Toff's research argues, it is ever more important for newsrooms that use AI in public-facing aspects of their journalism to do so with transparency and humility — with a focus on maintaining their audience’s increasingly diminished trust.
To that end, Walsh shared resources her organization offers newsrooms on how to talk with audiences about using AI, labeling content and conducting listening sessions to better understand audience needs and expectations around disclosure. Mahadevan stressed the importance of developing an organization-wide ethics policy around the use of AI — and shared the starter kit that Poynter put together last year to help guide that process.
Teams fan out in Murphy Hall to develop their prototypes
The Hackathon kicked off on campus Friday evening, Jan. 31, with a mixer and a brief presentation from the organizers about the ground rules and the ideas that came out of the MNA workshop earlier in the week. On Saturday morning, teams came together around the pitches they were most passionate about, and got to work.
By that afternoon, the eleven teams that succeeded in developing (mostly working) AI-powered prototypes presented their tools to their fellow participants and the panel of judges, which included representatives from Hacks/Hackers, the Minnesota Journalism Center and Poynter. Judges weighed a combination of six criteria for each project:
- Demonstrates clear potential to solve a real community need
- Shows a viable path to implementation beyond prototype stage
- Can scale to serve multiple communities or use cases
- Effectively uses AI capabilities
- Clearly discloses AI usage and limitations
- Has an intuitive interface that requires minimal training
The winning team, MinneDigest, will receive their award from Hacks/Hackers, which will serve as fiscal sponsors for the award. The team agreed to improve upon their prototype over the next three months by working directly with at least one local news outlet in Minnesota to refine the idea. In addition, Mahadevan agreed to serve as an ethics coach as the team develops the next iteration of the tool.
The Minnesota Journalism Center plans to follow the MinneDigest team's progress over the coming months — and to continue to grow the broader community that grew over the hackathon weekend at the intersection of technology, journalism and members of the community who are interested in addressing broader civic information needs.
As we build that community, we want to hear from people interested in joining us — in helping to develop the projects that teams began during our hackathon (full listbelow), and in connecting to journalists, tech builders or others engaged in the work of civic information. Join us here.
Final hackathon entries
Atticus: AI-Driven Local News & Civic Engagement Platform
Jason Kelly, Louann Berglund, Lori Ryan, Atticus Marse, Anand Kannan, and Ray Lewis
Atticus is revolutionizing local journalism by transforming complex civic data, government records, and hyper-local news into clear, accessible, and actionable content. Powered by AI-driven transparency, automation, and community collaboration, Atticus ensures that every Minnesotan—regardless of language, location, or digital access—stays informed and engaged with the decisions that shape their future.
BTTD (Belonging to the day)
Jon Vang, Josef Siebert
A tool to help local reporters and journalists discover story leads by monitoring digital platforms and allowing users to also submit ideas and have AI vet the story/submitter.
Frost Bite News
Shokria Ebadi, Allie Lopez, Hiba Harrari, Fabian Andrade, Bamlak Amedie, Maria Amedie
An AI chatbot to help journalists explore new ways to share news. While traditional media relied on articles and TV, today’s news spreads through social media and digital platforms. This chatbot transforms news into various formats like podcasts and short videos. It also connects to a customized system that prioritizes reliable sources and tailors updates based on location, allowing users to verify information as needed.
Legalese Simplifier
Janani Kannan, Priya Sisodia, Aishwarya Ramkumar, Rishab Agarwal
A tool to simplify legal texts, documents, and things like the “terms and conditions” when downloading apps online.
Local Government Bot
Chelsey Perkins, Megan Buffington, Jackie Renzetti, Jenifer Shier, Jeff Young, Sam Anklesaria, Wren Warne-Jacobsen, Kellen Baker
An aggregator that finds and summarizes local government agendas, minutes, and other documents and creates a centralized repository for them. Journalists or the public could enter their address and be given summaries of what their local governments are doing, plus directions on how to access the original documents. Journalists could use this as a central location for story idea generation, providing context and data analysis.
Localized Text Updates
Matt Mikus, Anish Saraf, Joshua O’Neill, Shourav Dey, Cynthia Tu, Eve Vang
Locally-tailored text messaging distribution platform for personalized alerts about weather or important government reporting.
M2F2 (Multi-modal Minnesota Fact Finding Service)
Shuning Lu, Hardik Gupta, Pravallika Kambhampati, Nick Hagar
A locally-grounded fact-checking app to verify statements made by political officials in Minnesota across different modalities.
MinneDigest: Your AI-Powered News Digest and Podcast
Dipan Bag, Arunachalam Manikandan, Erina Karati
AI-curated news summaries, personalized recommendations, and multilingual text-to-speech audio. Effortlessly browse and listen to local stories with smart insights and seamless accessibility. Your daily dose of Minnesota news—anytime, anywhere, in the language you prefer!
PDFtoText
Grant Hagstrom, Fred Olson
An easy to use tool that converts bulky pdf documents to usable text that does not rely on open source LLM tools but can be run entirely locally on journalists’ computers to preserve privacy and information security.
Simple English News
Yumiko Ehara, Jinny Zhang, Wenwen Cao, Jessica Tuleassi
A tool to translate local news into simple, easy, and accessible English, specifically for non-native English speakers.
VeriScore: Your Local Sports
Nick Capaldini, Anshu Patel, Michael Olson, Ningyuan Ma
A tool that collects the latest stats on high school and community sports into one dashboard, automates verification process with coaches via SMS text messages, and organizes content for local news outlets.
The Minnesota Journalism Center's work around AI and local news was made possible in part by support from a Rapid Response Seed Grant from the University of Minnesota’s Data Science Initiative. Hacks/Hackers also received some support from the MacArthur Foundation.
TOP PHOTO: Hackathon teams work in the Hubbard School's Murphy Hall on the second day of the Minnesota Journalism Center's AI + Journalism hackathon on Saturday, Feb. 1. Photo by Hannah Reynolds / HSJMC