Politics

Copy article

Reuters: Public Says "Not So Fast" on AI-Generated News

ended 08. October 2025

A new six-country survey by the Reuters Institute for the Study of Journalism  shows 61% of people now use AI systems (up from 40% in 2024), with weekly use jumping from 18% to 34%. Yet only 12% feel comfortable with fully AI-generated news, rising marginally to 21% with human oversight. In contrast, 62% trust entirely human-made content.

While AI's footprint expands (ChatGPT awareness hit 90%, and information-seeking via AI more than doubled to 24% weekly) trust lags behind utility. Just 6% now get news via AI systems (up from 3%), with younger users (8% of 18–24s) leading adoption. 

The comfort hierarchy is clear: People accept AI most where editorial judgment matters least. These are seen as admin tools: safe, invisible, and helpful.

  • 55% comfortable with grammar editing
  • 53% comfortable with translation

Moderate comfort is where there is some AI-heavy content published and confidence starts to wobble:

  • 30% comfortable with rewriting for different audiences
  •  27% comfortable with summarising/evaluating (48% among 18–24s)

But discomfort peaks where AI becomes the face or fabricates news content. Once AI becomes the content's face or voice, audience comfort evaporates:

  • 81% uncomfortable with artificial presenters
  • 74% uncomfortable with fabricated images replacing real photos

The oversight crisis: Only 33% globally believe journalists routinely check AI outputs before publishing, dropping to just 25% in the UK. This skepticism directly correlates with trust: 57% of those who strongly trust news believe checks happen, versus only 19% of strong distrusters.

The visibility problem: Despite 77% of respondents consuming news daily, only 19% see AI labelling. Yet Reuters' 2025 report shows 60% of newsrooms rate automation tools as "very important," in their workflows with 80% using AI for distribution and 77% for early content creation. The critical distinction: most AI use is backend productivity: transcription, tagging, translation, subtitles and not editorial judgment. The public's comfort levels suggest the answer: disclose when AI touches final published content, not the tasks mechanics.

The traffic threat: More people now see AI-generated summaries (54% weekly) than actively use AI tools (34%)—yet many don't click through to source links. This creates an existential challenge for news publishers who rely on search referrals: their journalism feeds the AI summaries that replace their traffic. 

As AI becomes the interface between readers and websites, publishers face a stark choice: adapt their business models or watch their content fuel systems that bypass their carefully curated news site entirely.

We'd like your views:

  • Can AI accelerate breaking news without sacrificing accuracy and will audiences trust it?
  • As 87% of media managers say generative AI has changed how their newsrooms operate, what editorial standards and training should become universal across the industry to ensure responsible implementation?
  • With 54% of people now seeing AI summaries weekly but many not clicking through to source articles, how can publishers maintain sustainable business models when their journalism feeds systems that bypass their website entirely in the user journey?

2 responses from the Newspage community

Copy all

Star Quote
Copy

Colette Mason
AI Ethics Consultant at Clever Clogs AI
Journalism is no different to any other industry grappling with AI integration. Speeding up the laborious production steps isn’t the problem. What is a problem is pretending you’re not doing it, while your readers are using the same tools daily to swiftly summarise documents, translate text, or check facts. The public gets it. They’re not scared of AI. They just don’t want to be lied to about where it is deployed responsibly.

Trust comes from being transparent about where AI is used, and proving that humans are still in charge of story accuracy.
Copy

Editor
Mitali Deypurkaystha
AI Governance Strategist & CMO at AISensa
AI may be rewriting how we find information, but it must not erase the people who create it. Google’s AI Overviews are fuelling a zero-click crisis, where readers get answers but journalists lose livelihoods. This isn’t innovation; it’s extraction. In an age already drowning in misinformation, we must protect reputable publishers doing the hard work of divulging the truth. Without fair compensation, independent journalism will collapse. If that happens, Google’s own summaries will choke on outdated content. Or, even worse, we get manipulated realities from corporations who don't have our best interests at heart. There’s already precedent: Canada's Online News Act has mandated that dominant tech platforms must compensate news organisations. Google has agreed to pay $100 million annually to Canadian news outlets. It needs to do this for news organisations worldwide. Google needs news the way humans need food. Paying for truth isn’t charity. It’s the nourishment of democracy itself.