USA Bureau

Leading US newsrooms are experimenting with generative artificial intelligence (AI) tools to enhance reader experience but a research reveals a significant gap between newsroom innovation and audience readiness.

A wide-ranging study by the Poynter Institute and the University of Minnesota indicates nearly half of Americans are not comfortable receiving news from generative AI, while one in five believe publishers should avoid AI entirely.

Dozens of America’s most well-known newsrooms including  the San Francisco Chronicle,the Texas Tribune, Time magazine and the Washington Post are experimenting with chatbots to help readers pick restaurants, learn more about political candidates and dive deeper into articles.

However, researchers suggest that public hesitation remains a key challenge.

Benjamin Toff, associate professor at the Hubbard School of Journalism and director of the Minnesota Journalism Center, presented the findings at the second Summit on AI, Ethics and Journalism, organised by Poynter and The Associated Press in New York City last week.

“The data suggests if you build it, do not expect overwhelming demand for it,” said Toff, who has been studying news audiences — and avoiders — for nearly a decade.

According to the Poynter Institute and the University of Minnesota, the survey found that 49.1% of respondents had no interest in using AI-based tools for information. Meanwhile, 39.3% said they would only use such a tool if editors verified its responses for factual accuracy, and just 9.9% expressed willingness to use the tools even if they occasionally misinterpreted published reporting.

Meredith Broussard, data journalist and associate professor at the Arthur L. Carter Journalism Institute of New York University, delivered a keynote at the summit where she spoke bluntly about user experiences with chatbots.

“Anybody really like using a chatbot? No. I can’t stand it. So, guess what? Your users feel like that, too,” she said. “They’re not excited about interacting with a chatbot on your site.”

The findings also revealed that many people have yet to interact with generative AI beyond customer service settings, making their scepticism about AI in journalism even more pronounced.

Furthermore, younger audiences, often perceived as early adopters of technology, are not as engaged with AI as expected. Nearly half of those aged 18 to 29 reported they hadn’t used or even heard of tools like ChatGPT.

Despite these reservations, some media organisations are pushing ahead with innovation. Hearst Newspapers launched the “Chowbot” in early 2024, an AI chatbot recommending restaurants based on decades of reporting.

Irish privacy regulator probes X over use of user data to train Grok AI

Ryan Serpico, deputy director of newsroom AI and automation at Hearst, defended the strategy, saying: “We are basing this off of 30 years of high-quality reporting, high-quality editing, that Google might push to the side or not value in their model.”

Christina Bruno, digital growth strategist at Spotlight PA, echoed the importance of exploring new formats. “We need to be experimenting with more formats of information delivery. I think chatbots are one way of doing that,” she said.

Internationally, audience-facing AI tools have gained more traction. In Sweden, publisher Aftonbladet’s EU election chatbot answered over 150,000 questions, while in Poland, a virtual assistant from Ringier Axel Springer helped generate 33,000 unique travel plans to promote German tourism.

“Experiments are great,” Broussard noted. “But you’ve got to pay attention to the results of the experiment.”

The study also showed a disconnect between public perception and newsroom practices. Respondents were asked, “Thinking about news media in general in the US right now, how often, if at all, do you think they currently use AI to do any of the following?” Of the 1,128 surveyed, 31.6% said AI is often used to make charts and infographics, and 6.2% said always.

For image creation when photographs are unavailable, 25.2% said often, and 6.2% said always. Meanwhile, 29.2% believed AI was often used to convert articles into audio or video, with 5.8% responding always.

Despite these assumptions, trust remains a pressing issue. More than half of respondents reported little or no confidence in newsrooms using AI to write articles or create imagery. For those with high news literacy, over 90% demanded clear disclosures when AI tools were used to generate text or edit photos.

Zuri Berry, digital strategy editor at The Baltimore Banner, sees this as a validation of their cautious approach. “It also serves as a confirmation of our current approach to AI, which entails disclosures, human review and verification and limitations on some tools that undermine our trust and credibility with readers,” he said.