From story ideas to deepfakes: How AI is remaking the newsroom

Artificial intelligence is moving into every part of news production, from finding leads to distributing stories. But its benefits depend on strong human oversight, according to Seth C. Lewis, who writes for Nieman Lab about the book Journalism in the Age of AI, co-authored with Rodrigo Zamith and Tomás Dodds.

The authors identify five areas where AI is already changing journalism: story ideation, sourcing, verification, storytelling and distribution. Tools can scan public records, social media and competitor coverage to identify possible stories. They can also transcribe interviews, search meeting recordings and help reporters examine large document collections.

Some applications can make reporting more inclusive. AI-powered sourcing audits can flag an overreliance on official voices and suggest experts or communities that are missing from a draft. Yet the authors warn that systems tend to find people with a strong digital presence. That can further exclude marginalized sources who lack institutional profiles or online visibility.

Speed helps, judgment remains essential

Verification is one of AI’s most promising uses. Fact-checking organizations use tools to detect claims in video and social posts, group recurring narratives and identify potential deepfakes. This allows journalists to focus their attention where it is most needed.

However, AI cannot reliably resolve claims when records are unavailable, sources refuse to respond or context determines whether a statement misleads. Large language models can also invent information. The authors argue that journalists must check the output of their own verification tools. Learn here how you can spot and avoid these “hallucinations”.

Generative AI is also producing headlines, summaries, translations and platform-specific versions of stories. It can help journalists build simple interactive tools without traditional programming skills. But recent cases involving fabricated bylines, invented quotes and inaccurate AI notifications show the reputational risks. The book’s central conclusion is that AI should augment newsroom work, not replace editorial responsibility, empathy or accountability.