Indianapolis newsrooms experiment cautiously with AI
Indianapolis newsrooms are testing artificial intelligence for specific tasks-transcription, graphics, data analysis-while maintaining human oversight. Most outlets have adopted policies that restrict AI to support roles rather than core reporting work.
The approaches vary widely. Black Indy Live has been the most experimental, using AI for logo design, graphics, story polishing, and promotional campaigns. The outlet even deployed an AI avatar named Rae to report news when it launched in 2017, though that ended when OpenAI discontinued the underlying technology.
Other local newsrooms take a more restrained stance. IndyStar requires strict transparency and human oversight. WFYI uses AI for transcription and captions but follows a "person-first/person-last" policy: a reporter starts the work and reviews it before publication. Indiana Capital Chronicle transcribes recordings with AI but has reporters verify quotes manually.
Mirror Indy transcribes notes and is developing a broader policy. Chalkbeat Indiana transcribes some meetings while attending key events in person. Axios Indianapolis, as a national company, has a formal partnership with OpenAI's ChatGPT for data analysis and document review. FOX59 and CBS4, both Nexstar stations, require that all writing and reporting remain the work of human journalists, with any AI use disclosed transparently.
What experts say journalists should do
Transcription tools like Otter are the most widely adopted AI application in newsrooms. Chatbots including ChatGPT, Claude, and Google's Gemini are gaining traction. Google's NotebookLM is emerging as a newsroom tool for organizing notes, interviews, and documents on individual stories.
Benjamin Toff, an associate professor at the University of Minnesota's Hubbard School of Journalism, and Alex Mahadevan, who trains journalists in AI at the Poynter Institute, agree on core principles. AI should serve reporting, not replace it. Using AI to generate entire articles or create fake images or videos crosses an ethical line.
Mahadevan flagged one specific concern: AI avatars designed to mimic humans. "When it comes to using AI to mimic a human being, that is a bridge too far and is ethically fraught because you are essentially, even if you are very clear with the disclosure, someone is still going to think that's a real human being," he said.
Good uses of AI include translating stories into different languages, converting text to audio for accessibility, and analyzing data for investigations. WFYI provides audio versions of stories. The New York Times and Washington Post already do this.
How to evaluate news sources using AI
Look for organizations that disclose how they use AI. WRTV and WISH-TV post disclosures at the end of stories. Check whether the disclosure is specific-mentioning the editorial team and standards for accuracy-or vague.
Verify that a human reviews AI work before publication. WFYI News Director Sarah Neal-Estes explained: "We don't just let AI do the work, we use it as a tool. For example, I do the interview first, person-first. I put it in Otter, then I fix that transcript, person-last; Otter always has mistakes."
Focus on a few trusted news sources rather than reading multiple versions of the same story. Toff said: "There are still some very good, in my view, quality news organizations that are investing in original reporting and news gathering, confirming information and doing the hard work that is journalism."
Be skeptical of AI used to create false content-photos, videos, or fabricated reporters. These deceive audiences regardless of disclosure.
What newsrooms should do next
Newsroom leaders should develop clear policies that staff understand and audiences can access. Mirror Indy created a committee to examine AI and is planning staff training. Chalkbeat has internal guidelines encouraging innovation while maintaining editorial standards.
Consider Prompt Engineering Courses for journalists who will use these tools regularly. Understanding how to ask AI systems the right questions improves output quality and reduces errors.
The stakes are clear: readers distrust AI in news for good reason. Newsrooms that use these tools transparently, keep humans in control, and focus on original reporting will stand out as the noise from low-quality AI-generated content increases.
Your membership also unlocks:
