Andrew Deck

News media is “becoming part of AI systems”: Notes from the JournalismAI Festival 2025 The London conference tackled topics like AI-assisted reporting, article personalization, and deepfake detection.

I was in London last month for the first in-person JournalismAI Festival hosted by @polislse.bsky.social and the Google News Initiative. I wrote up a few of the highlights for @niemanlab.org.
www.niemanlab.org/2025/12/news...

Politico management violated key AI adoption safeguards, arbitrator finds The ruling marks one of the first major tests of an AI clause in a newsroom union contract.

NEW: Last week an arbitrator ruled that Politico management violated key AI terms in its union contract.

The decision is a landmark win for the Politico union, which is one of the first to legally challenge the rollout of AI technology in its newsroom. www.niemanlab.org/2025/12/poli...

Florida nonprofit news reporters ask board to investigate their editor’s AI use Suncoast Searchlight’s four reporters told the board their editor-in-chief was using AI editing tools and inserting hallucinations into drafts. The next day, one of the reporters was fired.

Last week, Suncoast Searchlight's four reporters wrote a letter to their board. Their editor-in-chief had secretly been using ChatGPT and inserting hallucinated quotes into their drafts.

The next day, one of those reporters was fired. My latest for @niemanlab.org. www.niemanlab.org/2025/11/flor...

Apple Pulls China’s Top Gay Dating Apps After Government Order The removal of Blued and Finka marks another setback for China’s marginalized LGBTQ+ community.

SCOOP: China's top 2 gay dating apps, Blued and Finka, became unavailable on app stores over the weekend. Apple confirmed to WIRED that it removed the two apps following "an order from the Cyberspace Administration of China."

latest with @lmatsakis.bsky.social

Audiences doubt the benefits of AI-generated imagery in news are worth the risks, new study finds What do readers really think of AI-generated images in their news? There has been substantial research on how audiences respond to AI-generated text, but far less dedicated research on image generator...

✨ Audiences doubt the benefits of AI-generated imagery in news are worth the risks, new study finds.

@andrewdeck.bsky.social wrote an article on our latest study in @digitaljournalism.bsky.social 🙏

w Dina Strikovic

www.niemanlab.org/2025/11/audi... @aimediademlab.bsky.social @ddc-sdu.bsky.social

The Nonprofit Feeding the Entire Internet to AI Companies Common Crawl claims to provide a public benefit, but it lies to publishers about its activities.

NEW: Common Crawl, the massive archiver of the web, has gotten cozy with AI companies and is providing paywalled articles for training data. They’re also lying to publishers who have asked for material to be removed. “The robots are people too,” CC’s exec director told us when we asked about this.

Hundreds of thousands of videos from news publishers like The New York Times and Vox were used to train AI models YouTube channels from major news publishers and creators were in video data sets used by Microsoft, Meta, Snap, Runway, and Bytedance.

Grateful to the always-fantastic @andrewdeck.bsky.social for writing such a substantial piece building on our recent AI Watchdog work at the Atlantic 🙏

Tens of thousands of videos were also taken from some of the most popular news creators on YouTube.

“When the use is this concentrated it feels less like incidental inclusion and more like large-scale extraction of intellectual property," political talk show host @davidpakman.bsky.social told me.

AI Is Coming for YouTube Creators At least 15 million videos have been snatched by tech companies.

Using a searchable database published by The Atlantic, I identified downloads from specific news channels in video training data sets. These data sets were compiled or used by companies including Microsoft, Meta, Snap, Tencent, Runway, and ByteDance.
www.theatlantic.com/technology/a...

Hundreds of thousands of videos from news publishers like The New York Times and Vox were used to train AI models YouTube channels from major news publishers and creators were in video data sets used by Microsoft, Meta, Snap, Runway AI, and Bytedance.

Hundreds of thousands of videos were taken from some of the most popular news publishers on YouTube and used to train AI video models. That includes The New York Times, ABC News, Al Jazeera, Vox Media, and The Wall Street Journal. My latest @niemanlab.org. www.niemanlab.org/2025/10/hund...

The Wayback Machine’s snapshots of news homepages plummet after a “breakdown” in archiving projects Between May and October 2025, homepage snapshots fell by 87% across 100 news publications.

The Wayback Machine’s snapshots of news homepages have dropped off in recent months. Our analysis shows that since May 16, the number of captures available for 100 top news sites declined by 87%.

@hanaatameez.bsky.social and I report for @niemanlab.org.
www.niemanlab.org/2025/10/the-...

Scammers are using video deepfakes of journalists to peddle products online Broadcast journalists around the world are being targeted amid an explosion in deepfake scams. Are social media platforms doing enough to counter them?

Deepfakes of broadcast journalists are spreading across social media. Many of them are ads selling cryptocurrencies, gambling apps, and snake oil medicines.

Meanwhile, social media companies like Meta profit off the ad dollars. My latest @niemanlab.org. www.niemanlab.org/2025/10/scam...

A Journey Into the Heart of Labubu I made an epic trek across four countries to answer one question: Why is the world going mad for a plushie monster?

In college, I'd always walk past an unremarkable store in Beijing that sold a variety of trendy, but tacky, toys. That store has turned into Pop Mart, the $45-billion Chinese toy empire behind Labubu, the biggest global fad of 2025.

This is my big story/pilgrimage trip to explain how they got here

Tech Billionaires Already Captured the White House. They Still Want to Be Kings From Montenegro to Northern California, the tech elite dream of building cities where they make the rules. Is this, finally, their moment?

NEW from me for @wired.com's politics issue: I have spent SO much time fascinated by the "network state"--privatized cities in special economic zones (SEZs) where the normal rules and regulations don't apply.

www.wired.com/story/tech-b...

Local LGBTQ+ publications are facing a drop in ad dollars and a rise in safety concerns, new report finds A new report from News is Out, a national collaborative of queer news publishers, has mapped local LGBTQ+ publications across the U.S. The first comprehensive report of its kind, “The LGBTQ+ Media Map...

"I’ve had people tell me they're afraid to go on the air, and after 40 years on the air, that’s the first time I’ve ever heard that."

A new report from @newsisout.bsky.social asks how local LGBTQ+ news outlets are weathering the second Trump administration. www.niemanlab.org/2025/09/loca...

Nieman Lab’s Andrew Deck Has Been Tracking AI Use in the Journalism Industry for Years. Here’s What He Says We Need to Know. - Ethics and Journalism Since the public launch of ChatGPT nearly three years ago, the media industry has cycled through many phases of AI adoption and experimentation.

Great to chat with NYU’s Ethics & Journalism Initiative about my reporting!

We talked about AI "bias detectors,” automated fact-checking, and navigating both the hype and fear cycles on my beat. ethicsandjournalism.org/2025/09/18/n...

The True Threat of OpenAI Karen Hao’s recent book on the company argues that its ambitions are not merely about scale or the market but the creation of a global force that rivals a colonial power of old.

For @thenation.com: I reviewed @karenhao.bsky.social’s Empire of AI and took a closer look at her argument that OpenAI’s "scaling laws" echo the expansionist ideologies of European empires.
www.thenation.com/article/cult...

Business Insider and the Bots The Axel Springer-owned newsroom is buzzing over new ChatGPT writing guidelines—part of an aggressive A.I. strategy pushed by its German parentco and detailed in a memo obtained by Status.

Scoop: Business Insider informed its staff this week that they are allowed to use ChatGPT to generate first drafts of their stories, while also indicating the newsroom will not disclose such A.I. use to readers.

Details in @status.news: www.status.news/p/business-i...

From reckoning to retreat: Journalism’s DEI efforts are in decline Diversity-related newsroom jobs haven't totally disappeared — but they also haven't stuck.

A third of jobs related to race, diversity, and equality that were created in U.S. newsrooms since 2020 are gone as of this summer. 

@hanaatameez.bsky.social with a months-long investigation and new data on American journalism's turn away from DEI initiatives. www.niemanlab.org/2025/09/from...

AI Watchdog landing page

AI Watchdog Data Set Search

I'm proud to share a new project that @alexreisner.bsky.social and I launched at @theatlantic.com today! It's called AI Watchdog, and it's our new home for all of the investigations into training data sets, such as LibGen, Books3, and OpenSubtitles. www.theatlantic.com/category/ai-...

Some French publishers are giving AI revenue directly to journalists. Could that ever happen in the U.S.? Le Monde agreed to give journalists 25% of revenue from licensing deals with OpenAI and Perplexity. Now, other French publishers are following suit.

In France, reporters are getting a cut of revenue from AI licensing deals. Among others, Le Monde has agreed to redistribute 25% of its deals with OpenAI and Perplexity.

I took a closer look at these union agreements and asked why they haven't crossed the Atlantic. www.niemanlab.org/2025/09/in-f...

NEW in @columjournreview.bsky.social: Why AI models are bad at fact-checking photos www.cjr.org/tow_center/w...

Local newsrooms are building AI chatbots fast and cheap A new report tracked four small newsrooms as they launched custom chatbots built in just one month.

A new report from @centerforlocalnews.bsky.social asks what it takes for small local newsrooms to build an AI chatbot.

Researchers helped four newsrooms launch experimental chatbots in under a month. But only one has committed to running theirs beyond the program. www.niemanlab.org/2025/08/loca...

How tribal radio stations are preparing for a future without the Corporation for Public Broadcasting “These stations are not just media outlets, they are cultural infrastructure.”

Quite the line in this excellent @niemanlab.org piece: "These stations are not just media outlets; they are cultural infrastructure." www.niemanlab.org/2025/08/how-...

Japan’s largest newspaper, Yomiuri Shimbun, sues AI startup Perplexity for copyright violations The Yomiuri Shimbun, Japan’s largest newspaper by circulation, has sued the generative AI startup Perplexity for copyright infringement. The lawsuit, filed in Tokyo District Court on August 7, marks t...

Yomiuri Shimbun, Japan’s largest newspaper by circulation, has sued Perplexity for copyright infringement.

The filing claims that Perplexity accessed 119,467 articles on Yomiuri’s site between February and June of this year.
www.niemanlab.org/2025/08/japa...

Politico’s recent AI experiments shouldn’t be subject to newsroom editorial standards, its editors testify In a July arbitration hearing, Politico faced allegations that two generative AI tools violated its union contract.

In a July arbitration hearing, Politico's top editors testified that two recent AI experiments don't need to meet journalistic editorial standards.

Union members alleged those experiments violated AI guidelines in their contract. My latest @niemanlab.org.
www.niemanlab.org/2025/08/poli...

Generative AI models love to cite Reuters and Axios, study finds Muck Rack found that major models cite journalism in nearly half of responses that require "recency."

A new study by Muck Rack recorded over 1 million citations output by major generative AI models.

Journalistic content was cited in 27% of responses, a number that jumped to 49% when those responses required "a level of recency." www.niemanlab.org/2025/07/gene...

Inside a network of AI-generated newsletters targeting “small town America” Good Daily, which operates in 47 states and 355 towns and cities across the U.S., is run by one person.

Here's our original story on Good Daily: www.niemanlab.org/2025/01/insi...

And a database I compiled earlier this year, which lists 337 of its local newsletters: www.niemanlab.org/2025/02/the-...

6AM City acquires Good Daily’s network of more than 350 AI-generated local newsletters 6AM City will continue to operate its "core" newsletters with human editors, but will treat Good Daily’s AI-generated newsletters as "seed markets."

In January, I uncovered a network of AI-generated newsletters, called Good Daily, that was targeting hundreds of small towns and cities across the U.S.

This week, the newsletter company 6AM City announced it had bought Good Daily and brought its founder on board. www.niemanlab.org/2025/07/good...

What went wrong at the Houston Landing? “We tried to be too much, too fast, for too many people.”

If you are involved in trying to make nonprofit news sustainable, I wrote and reported this story for you!

So much news that deserves our attention right now, but I hope you'll spend some time with my Houston Landing retrospective for @niemanlab.org (1/3)

www.niemanlab.org/2025/07/what...