By Jessica Mahone, Interim Director, CISLM
At the Center for Innovation and Sustainability in Local Media, we spend a lot of time thinking about what makes local journalism work — and what our highest impact might be on making it work better for audiences, journalists themselves and news organizations as a whole.
Lately, that’s meant looking closely at artificial intelligence. Not because it’s trendy or futuristic, but because it’s here. And local news organizations, like the rest of us, are trying to figure out what role it might play in their work. To that end, journalism support organizations and funders alike are trying to establish programs to help news orgs experiment with AI, but not many results of such programs are public.
Our newest research experiment is a model we’re proud of: A hands-on, newsroom-centered pilot we’re calling the Local NewsBot Studio, combining product development with industry-focused research and insights.
Our pilot project focuses on chatbots for local news organizations, custom built with newsrooms, with their audiences in mind. With it, we’re exploring how conversational interfaces can support community engagement, reduce inefficiency in staff time, increase visibility for existing reporting and even tie into revenue goals like newsletter growth and subscriptions.
What the Local NewsBot Studio Is
The Local NewsBot Studio is a collaborative program where we at CISLM work side-by-side with four newsrooms across the Southeast — spanning community radio, digital and a legacy news organization — to build custom AI-powered chatbots. Each bot is designed around a specific audience need or newsroom opportunity, using the newsroom’s own information as its knowledge base.
We’re grateful to partner with four wonderful organizations to support the following projects:
- Ask ACC, an assistant to help with questions related to civil politics and civic life in Atlanta, especially designed with the upcoming elections in mind. This bot is built for Atlanta Civic Circle, a nonpartisan, nonprofit civic journalism platform designed to inform and inspire metro Atlanta residents to actively participate in addressing the most important issues facing our region.
- Chappy, the personal assistant for Chapelboro. In the style of Clippy™, Chappy helps readers navigate news from Chapelboro and 97.9 The Hill, summarizing info from the past year and answering frequently asked questions about the organization. Digital site Chapelboro and radio station 97.9 The Hill are combined on this site to provide total community coverage in Orange County, NC.
- Henry, the personal assistant for Henrico Citizen in Henrico County, VA. Also in the style of Clippy™, a personal assistant for readers to navigate the Henrico Citizen website by answer frequently asked questions about the site, including how to find stories, navigate articles, or any other inquiries, and finding and synthesizing news stories published in the past year specifically related to Henrico County.
- News Reporter Help Desk, the front office bot for the News Reporter in Whiteville, NC. The small editorial staff of the Pulitzer Prize-winning newspaper was spending a large chunk of their time answering the same questions about membership, submitting stories and obituaries, placing ads and more. This bot helps to answer “front office” questions about how readers can engage with the paper.
We’re providing everything from the technical lift (consulting with news organizations, building and helping to embed the chatbot) to training and audience data support.
We’re building on what we learned from our own AI-powered chatbot, the Local News Audience Assistant, with a unique team with backgrounds in product and research that makes this possible:
- Local News Researcher Yanan Sun is a former entrepreneurial fellow at the Brown Institute for Media Innovation and has co-founded an AI product that helps users read mobile apps’ privacy policies;
- Project Manager Sarah Vassello built our Local News Audience Assistant chatbot for CISLM, which compiles case studies, best practices, Q&As, tip sheets, audio and video recordings and more around audience development from journalism support organizations.
- Interim Director Jessica Mahone is a celebrated researcher in local news.
Our goal is to help these organizations test real applications of AI — while giving us a clearer understanding of how people engage with these tools in a local news context.
Why chatbots?
A few reasons:
- They meet user needs in a variety of ways. Chatbots offer an easy, intuitive way for people to find relevant information — especially when designed with care and specificity. This can be true to find certain content items or to help readers engage with the business side of an organization.
- They can surface valuable reporting. Many local newsrooms have deep archives of community information. Chatbots can make that more accessible.
- They can save time and allow newsrooms to focus on reporting. Many local newsrooms have a small staff who do everything from reporting to sales to customer service. Chatbots can fill customer service and other business needs, allowing staff to prioritize meeting their communities’ information needs.
- They give us new signals. Through user queries and feedback, newsrooms can learn more about what their audiences are looking for.
We’re not just interested in whether these tools work — we want to understand how news audiences would like to engage with chatbots, and thus AI, and why. That’s the research question behind this project.
Keeping It Human — and Local
AI has well-documented pitfalls. One of the most important parts of this work is making sure we don’t sacrifice trust or accuracy in the name of automation. That’s why our chatbots are powered by hand-curated spreadsheets of local reporting — not open web scraping or general knowledge models. And it’s why both our team and participating newsrooms will monitor chatbot conversations and address any issues along the way.
We’re also thinking about sustainability. At the end of the program, our partners will walk away with their own chatbot, analytics, and training — equipped to decide if and how they want to continue using the tool after the pilot ends.
What Comes Next
We will collect data from the bots through early June. After that, our research team will dive into the data, talk with participants, and publish our findings in late June or early July. We’ll be asking: What worked? What didn’t? And what can other local newsrooms learn from this?
We also hope to provide findings to funders and support organizations with AI programming to better support productive programs for news organizations.
This project is as much about learning as it is about building. Let us know your thoughts about AI, product development/research opportunities and more.
—
Curious about the process? We’ve documented our build step-by-step here.
Questions? Reach out to me anytime at jamahone@email.unc.edu.