A new report from university researchers reveals that artificial intelligence is already a significant presence in American journalism, with an automated analysis of 186,000 recent articles concluding that about 9% are at least partly written by machine. This growing integration of AI into news production is happening largely without public disclosure, creating a gap between newsroom practices and audience awareness. The study found that this AI usage is not uniform, appearing more frequently in resource-strapped local publications and in specific sections like weather and health, raising fundamental questions about transparency and the future of journalistic labor.
While the technology offers a way for smaller newsrooms to fill pages amid shrinking budgets, its application is often hidden from readers. The academic report notes that AI use is “widespread, uneven, and rarely disclosed,” leaving communities who depend on local outlets for civic information unaware of the software’s role in shaping the news they consume. This quiet adoption runs parallel to widespread public skepticism, with separate polling data showing a majority of Americans are wary of AI’s role in journalism and concerned about its potential to spread misinformation.
Prevalence in News Reporting
An extensive academic analysis of 186,000 articles across a wide range of U.S. newspapers showed that approximately 9% of newly published stories were either partially or fully generated by AI. The researchers utilized an automated detection system to identify machine-created text, uncovering a significant and previously unquantified level of AI involvement in daily news production. The report highlights that certain topics are more susceptible to automation, as they often rely on standardized information and clear numerical data that can be easily fed into writing templates. Among the news categories with the highest rates of AI-detected content were weather, science, technology, and health.
The study also points to disparities in language, noting that Spanish-language articles published by American newspapers contained more automated text than their English-language counterparts. This finding suggests that the application of AI tools varies not only by topic but also by the target audience and the specific newsroom’s workflow and resources. The overall picture is one of a technology being integrated quietly but firmly into the fabric of American news gathering and dissemination.
Disparities in Publication Type
The use of artificial intelligence in news generation is not distributed evenly across the American media landscape. Instead, it appears to be strongly correlated with the size and circulation of the news organization. According to the study, AI-generated content is significantly more prevalent in local newspapers with smaller circulation numbers compared to major, nationally circulated papers. The researchers suggest that this trend is linked to the economic pressures facing smaller publications, which often operate with lean staffing and fewer resources. In these environments, automated tools can help fill content gaps and maintain a high volume of daily reporting that would otherwise overwhelm the available journalists.
As the report states, automation tends to rise where circulation drops. Larger, more established newspapers do not appear to be using AI in the same capacity for their core news reporting. This creates a divide in how news is produced, with readers of local papers being more likely to encounter machine-written articles without their knowledge. For many communities, these smaller outlets are the primary source of information about essential civic life, yet they are also the most likely to be supplementing human journalism with undisclosed AI.
AI in Opinion Content
The study revealed a particularly striking trend on the opinion pages of prominent newspapers, including The New York Times, The Washington Post, and The Wall Street Journal. In a departure from the patterns seen in general news, opinion content was found to be 6.4 times more likely to contain AI-generated material than news articles from the same publications. This surge in AI use on pages designed to shape national discourse on major issues has grown significantly since the end of 2022, coinciding with the widespread availability of advanced new writing systems.
The analysis further breaks down the source of this automated content, finding that guest contributors—those who are not permanent staff columnists—are more frequent users of AI writing assistance. This suggests that while established, in-house commentators may be writing their own material, external experts and writers contributing to the opinion section are increasingly relying on software tools. The finding complicates the traditional understanding of opinion journalism as a space for distinct human voices and perspectives, introducing a new layer of technological mediation that is largely invisible to the public.
Public Perception and Distrust
While some newsrooms adopt AI tools, the American public remains highly skeptical of their use in journalism. A poll conducted by Poynter and the University of Minnesota found that nearly half of American respondents do not want AI reporting the news, and 20% believe publishers should not be using the technology at all. The survey, which polled 1,128 people, revealed deep-seated distrust. For instance, 30% of those surveyed expressed “no confidence at all” in news outlets using AI to write articles, and 32% felt the same about AI creating images where real photographs are unavailable.
This sentiment is echoed by a Pew Research Center survey, which found that roughly half of U.S. adults believe AI will have a negative impact on the news people receive over the next two decades. A primary concern among the public is the potential for inaccurate information, with 66% of Americans stating they are extremely or very concerned about this possibility. Furthermore, there is widespread belief that AI will negatively affect the journalism profession itself; 59% of Americans think the technology will lead to a reduction in jobs for journalists in the coming years. Despite this wariness, broader societal use of generative AI for information-seeking has nearly doubled in the past year, indicating a complex and evolving relationship with the technology.