One of the better newsroom AI use cases:
POLITICO consolidated 6 CMSs into one system and can now use AI to help analyze 20,000 ICE detention rulings, monitor legislative data and reduce workflow friction.
https://t.co/B1RD5aqKlX
Interesting lesson from FOIAball:
A paywalled scoop produced 1 new paid subscriber.
A free scoop produced 3.
People often don’t subscribe because one article is locked away.
https://t.co/0JzV0CU7gX
AI monitoring public meetings and helping local reporters find stories? Useful.
AI sites repackaging existing local journalism and calling it “rebuilding local news”? That’s extraction, not innovation.
https://t.co/qTzE0uhr0L
AI helping research or fact-check an op-ed? Fine.
AI substantially writing the op-ed? That deserves disclosure.
A byline is supposed to mean more than “this person agrees with the output.”
https://t.co/i1cODBIfpm
Daily Mail, BBC, CNN and Fox News have each added 11M+ TikTok followers since 2024.
Young audiences clearly haven’t stopped caring about news. They’re just consuming it differently.
Newsrooms should learn vertical video and platform-native storytelling.
https://t.co/QzfErEy1DS
The best part of Google’s Backstory isn’t AI detection.
It’s context.
Reverse-image search, provenance, watermarks + a timeline of where an image appeared can tell journalists far more than a black-box “87% AI” score.
And it still gets things wrong.
https://t.co/7HOLNY7lhR
Newsrooms don’t need to turn every reporter into an influencer.
But they should learn from creators.
The Athletic’s model makes sense: creators bring personality, platform fluency + audience instinct; journalists bring reporting, editing + fact-checking.
https://t.co/bKbKFDkkUO
Prism News built 200+ AI “local news” sites, repackaged reporting from real outlets under fictional bylines, produced no original journalism — and now the network is gone.
A news desert is not a shortage of websites.
https://t.co/0V29gCmzML
Reddit turning posts into AI-narrated videos could be useful for accessibility.
But Reddit’s real value isn’t the text. It’s knowing there are actual people behind it.
https://t.co/vKLXu5CDos
AI helping a local reporter draft records requests and catch missing government documents? Great.
AI helping justify one reporter covering two counties larger than Rhode Island? Much less convincing.
That’s the tension in Axios’s bet.
https://t.co/nbmraMOTUx
Apple reportedly wants to pay publishers when Siri AI actually uses their journalism.
Good principle: if reporting creates value for an AI answer, the newsroom should share in that value.
https://t.co/m5IVBCbQpM
What if public media built social media around civic value instead of engagement?
https://t.co/M7alx60FnX is trying: no follower counts, chronological feeds, pro-social reactions + paid human community stewards.
https://t.co/nRF2fxJuoj
Ad tech isn’t just about targeting ads.
The same real-time bidding infrastructure can expose where your phone is — and that location data can end up with brokers or government customers.
Tools like DecryptAds help make the system visible.
https://t.co/BLVY0P3PId
Fake sites copying major news brands, filling them with AI-generated PR, then selling “guaranteed coverage” aren’t journalism.
They’re counterfeiting journalism’s credibility.
Good PR can pitch a story. It cannot guarantee independent coverage.
https://t.co/vs1uL5vNuR
Maybe we don’t need another algorithm claiming to tell us what’s “true.”
Japan’s Originator Profile takes a smarter approach: cryptographically verify who actually published something.
It won’t replace fact-checking.
https://t.co/QbzbpGHY58
Publisher-owned AI is a lot more interesting than outside platforms summarizing journalism and keeping the audience.
The NY Post’s Hamilton uses its own reporting for search and personalization while retaining editorial control.
https://t.co/xIn2YCvWDj
More comments don’t necessarily mean better engagement.
A study of nearly 1M NYT comments found that data-rich, analytical stories drew fewer commenters—but more evidence-based arguments and reader-to-reader discussion.
https://t.co/F0XL3DSxkh
SPJ is updating its ethics code for the AI era without writing rules around any specific AI tool.
That’s the right approach.
Whatever technology journalists use, humans still own the accuracy, verification and consequences.
https://t.co/5QmJwQ62fi