UW’s Public Tech Media Lab helps bridge the gap between AI and journalism

The field of journalism, and media in general, is facing a pivotal moment. The rise of artificial intelligence technologies is changing the way journalists report on some of the most pressing issues. But while AI remains controversial, it also has the potential to be a valuable tool for newsrooms, provided those working in the media landscape are equipped to use it responsibly and hold it accountable.

One of the biggest challenges facing newsrooms with AI is the pace at which new technologies are entering the field, which may prompt media companies to implement tools quickly - such as large language models (LLMs) to aid reporting - without being intentional about what they use and how.

Assistant Professor Tomás Dodds , a new RISE-AI faculty member in the School of Journalism and Mass Communication, aims to bridge the gap between the world of AI and journalism. He hopes to not only research the impact one may have on the other, but also to shape the conversation around how AI could be used to enhance reporting and storytelling.

"Journalism studies tend to be very extractivist, which means that we usually interview journalists, then write papers that not a lot of people are going to read," Dodds explains. "Based on the Wisconsin Idea, I felt that we were missing a space where we could bring together practitioners, journalists, academics and researchers to help newsrooms make responsible technological decisions. That is what the Public Tech Media Lab is."

The Public Tech Media Lab has already partnered with local newsrooms (e.g., Isthmus , PBS Wisconsin and Wisconsin Public Radio) to identify AI guidelines and adopt AI systems that support their work while adhering to core reporting values and standards.

These AI systems could include in-house LLMs or ethical AI-assisted recording platforms that journalists can trust with sensitive or private information. This could also reduce the dependence of some media companies on large technology platforms, such as OpenAI's ChatGPT, Microsoft Copilot or Otter.ai, where data or personal information may be used in ways a journalist did not intend.

The lab recently received funding approval for a postdoctoral position to develop AI tools for newsrooms.

For example, a postdoc could help newsrooms develop an in-house LLM trained solely on their own archival data, encouraging reporters and editors to interact with previous work, a resource that newsrooms may overlook when reporting new stories.

"We want to reach as many newsrooms as we can to show media organizations that they can have in-house solutions that adapt to their ideas and values, and can be modified depending on their needs," Dodds says. "A lot of media organizations cannot do it themselves, because it has an economic cost, but also, more than anything else, it requires a lot of technical knowledge. That's why the lab is here. It is an amalgamation of computer scientists, lawyers and journalists."

While one of the lab's goals is to support local newsrooms through discussions and the adoption of AI tools, Dodds says they also aim to expand these resources beyond Madison to benefit academia and media organizations globally.

All resources and tools developed by the lab will have open access, meaning organizations on the other side of the world could also access data sets, tools or systems.

Not only will this approach support media organizations beyond Wisconsin that may lack funds or resources to incorporate ethical AI tools, but it will also ensure transparency regarding the lab's work.

"The idea of the lab is not just to generate academic knowledge, but also to make that knowledge actionable for academics and media organizations around the world, and that's really important," Dodds explains. "If we conduct a survey, people will be able to access that data. If you're a newsroom or an academic from anywhere in the world that may not have access to these funds, you can access the materials we create."

Dodds is also publishing a new book this August co-written with Seth C. Lewis, distinguished professor of artificial intelligence and media studies at the University of Virginia, and Rodrigo Zamith, the chair of the journalism department at the University of Massachusetts Amherst, titled Journalism in the Age of AI: From Acceleration to Reimagination . Thanks to the RISE Initiative, the book will be open access, making it available to students, journalists and academics everywhere.

The lab also aims to provide additional opportunities to educate journalists on open-source investigations, including how information circulates and how to employ data-collection strategies to leverage publicly available sources. In late May, the lab will host a pair of two-day Digital Investigations Bootcamps for journalists and editors on open-source intelligence (OSINT).

Beyond applying these tools and methods in newsrooms globally, Dodds seeks to bring the conversation to UW classrooms, with the goal of creating the first open-source intelligence course in the Midwest. The course would cover how journalists can use artificial intelligence in their work, positioning UW-Madison journalism students to navigate this complex topic as they enter professional careers.

"There are very few universities doing anything with open-source intelligence, so UW-Madison has the opportunity to be a pioneer on this particular topic," Dodds says.