Study finds 1 in 10 U.S. newspaper pieces partly AI-written, often undisclosed
When I looked at the latest study on U.S. print journalism, the headline that stuck was that about ten percent of articles now hide AI-written bits, and most readers never notice. The researchers actually sifted through thousands of pieces, from big national dailies down to small regional papers, and marked any passage that fit the tell-tale patterns of large language models.
What they found was a clear rise, especially on editorial and commentary pages where the voice can drift just enough to feel off. Newsrooms often brag about the speed gains, yet the same report points to a growing legal tug-of-war: a handful of top publishers have sued AI firms, saying those companies scraped their copy to train algorithms. Ironically, those very outlets are slipping AI-crafted prose into their own pages without much fanfare.
It’s hard to say how this will play out, but the mix of secrecy, copyright claims and the subtle shift in tone certainly puts journalistic integrity on shaky ground.
The study points out that several major media groups are suing AI companies for scraping their articles to train language models, even as those newsrooms quietly publish AI-generated stories themselves. The surge in AI-generated content is particularly notable in opinion sections of national newspapers. An analysis of 45,000 op-eds and commentaries from The New York Times, Washington Post, and Wall Street Journal between 2022 and 2025 shows a 25-fold increase in AI use--from just 0.1 percent to 3.4 percent. Opinion pages are now 6.4 times more likely to feature AI-written content than regular news stories at these publications.
It seems the public rarely gets a heads-up when a story has a machine-crafted component. A University of Maryland analysis puts the number at roughly one in ten U.S. newspaper pieces containing AI-generated text, and most readers never notice.
The researchers ran the Pangram detector on more than 250,000 recent articles and reported a false-positive rate of just 0.001 % for news content, which makes the figure look more than a statistical blip. Adoption, however, isn’t even. Local outlets appear far more likely to use the tech than big national papers, and opinion sections in major dailies are showing a noticeable uptick in AI-assisted copy.
At the same time, a handful of large media groups have sued AI firms for scraping their archives to train models, even as their own newsrooms slip AI-written material into the workflow. The study leaves it unclear whether this will chip away at trust or simply settle in as another editorial tool. Until we get clear disclosure rules, readers are left guessing about the human or algorithmic roots of the stories they read.
More in Research & Benchmarks
Dell and NVIDIA host AI developer meetup in Bengaluru on deployment trade‑offs
Dell and NVIDIA have converged on Bengaluru for a hands‑on gathering aimed at the engineers and...
GPT-5.2 leads FrontierScience test, but falters on real research tasks
OpenAI’s latest benchmark, the FrontierScience test, pits its newest model, GPT‑5.2, against a...
OpenUSD and NVIDIA Halos Enhance Robotaxi Safety with Synthetic Data, SimReady
Robotaxi developers have long wrestled with the gap between virtual testing and real‑world...
Audio Dataset Valuable for Listening Models, Tackles Noise, Accents, Timing
Why does a single audio collection matter when the field is flooded with text‑heavy benchmarks?
Fastweb and Vodafone use LangGraph LLM Compiler to automate customer requests
Why does this matter? Because two of Italy’s biggest telecoms have moved from isolated chatbots to...
Common Questions Answered
What proportion of U.S. newspaper articles were found to contain AI‑generated text in the University of Maryland study?
The study determined that about one in ten, or roughly 10 %, of U.S. newspaper pieces include passages generated by artificial intelligence. This figure comes from analyzing more than 250,000 recent articles across national and regional outlets.
How did researchers identify AI‑written passages, and what was the reported false‑positive rate?
Researchers applied the Pangram detector, a tool designed to spot patterns typical of large language models, to the article corpus. They reported a false‑positive rate of just 0.001 % for news content, indicating a high level of detection accuracy.
Which sections of newspapers saw the sharpest increase in AI‑generated content, according to the analysis of op‑eds and commentaries?
The surge was most pronounced in editorial and commentary pages, especially opinion sections of national newspapers like The New York Times, Washington Post, and Wall Street Journal. Between 2022 and 2025, AI use in these sections rose 25‑fold.
What legal actions are major media groups taking against AI companies, and how does this relate to their own use of AI‑generated stories?
Several large media groups are suing AI firms for scraping their articles to train language models, alleging copyright infringement. Paradoxically, many of these newsrooms are also publishing AI‑generated stories without disclosing the machine‑crafted portions to readers.