News trust hits record low, but newsrooms still matter

Key Takeaways

• Social media and video networks now surpass traditional sources for news consumption at 54%
• Global trust in news hit an all-time low of 37%, with US trust at just 25%
• AI chatbot news consumption grew from 7% to 10% in one year, signaling a shift toward synthesized content

News trust has fallen to its lowest point since Reuters began tracking it in 2015. Just 37% of people globally say they trust most news most of the time. In the United States, that figure drops to 25%. The data comes from the 15th annual Reuters Institute Digital News Report, which also marks a first: social media and video networks have overtaken traditional sources as the primary way people consume news.

Across audiences surveyed, 54% now get their news from social media and video platforms like YouTube. That edges out publisher websites (51%) and TV (52%). The shift represents a structural change in how information flows to audiences.

Why is news trust declining so sharply?

The Reuters data aligns with other long-running measures. Gallup's October 2025 survey put US trust in mass media at a record-low 28%, down from 31% the previous year and 40% five years ago. The decline is consistent across polling methodologies and time periods.

One factor accelerating the trend: AI chatbots. The Reuters report puts news consumption via AI chatbots at 10%, up from 7% a year ago. That's a 43% jump in twelve months. Chatbots synthesize content from multiple sources into summaries, stripping away the brand identity of the original publisher. The atomic unit of journalism shifts from the article to the underlying facts, quotes, and insights that an AI can extract and remix.

What does shallow consumption look like?

Most news consumption on social platforms is surface level. People scroll past news posts mixed with fitness tips, product ads, and entertainment content. They absorb headlines without clicking through. The Reuters report identifies a growing segment: 12% of people now get news only by bumping into it on social platforms while online for something else. That figure has doubled since 2020.

This creates a paradox. People consume more news fragments than ever, but they trust less of what they see. The feed environment conditions audiences to treat news as another content type competing for attention, not as a distinct category requiring different evaluation standards.

Do newsrooms still carry weight?

Here's where the data gets interesting. The title of the Fast Company analysis captures the nuance: people have stopped trusting news, but not necessarily newsrooms. When audiences click through to a publisher site, when they seek out a specific outlet, the relationship is different from passive feed consumption. The 51% still going directly to publisher websites represent an intentional choice.

The distinction matters for AI builders. Users may distrust "the news" as an abstract category while still valuing specific sources. Brand attribution in AI-generated summaries isn't just a legal question. It's a signal that affects how users weight information. An AI system that cites Reuters or the Financial Times triggers different credibility judgments than one that presents facts without provenance.

What happens when publishers become wholesalers?

The Reuters report points toward a future where publishers function as information wholesalers. They produce the raw material: original reporting, verified facts, expert quotes. AI systems then aggregate, interpret, and distribute that material in new forms. The publisher brand fades to the background.

This model has precedent. Wire services like AP and Reuters have always operated as wholesalers to other media outlets. The difference now is that the "retail" layer may be an AI chatbot rather than a newspaper or broadcast network. And unlike traditional syndication, the AI layer doesn't pay licensing fees by default.

Some publishers are negotiating deals. OpenAI has agreements with outlets including the Associated Press, Axel Springer, and Le Monde. Google's AI search summaries cite sources with links. But the economics remain unsettled, and smaller publishers lack leverage in these negotiations.

The trust gap as a product opportunity

For AI product teams, the trust collapse creates both risk and opportunity. Risk: if your AI system surfaces news content, users may distrust it by association. The 25% trust figure in the US means three-quarters of Americans approach news with skepticism. That skepticism transfers to any product that delivers news.

Opportunity: products that restore trust signals have a competitive edge. That might mean transparent sourcing, clear attribution chains, or user controls over which outlets an AI can draw from. Perplexity's recent "Pages" feature, which lets users see every source behind a summary, moves in this direction. So does the citation format in Claude's web search results.

Frequently Asked Questions

What percentage of people trust news in 2025?

Globally, 37% of people trust most news most of the time, according to the Reuters Institute Digital News Report. In the United States, the figure is 25%.

How many people get news from AI chatbots?

10% of people now consume news via AI chatbots, up from 7% the previous year, representing a 43% increase.

Has social media passed TV for news consumption?

Yes. 54% of people now get news from social media and video networks, surpassing TV (52%) and publisher websites (51%) for the first time.

What is the Reuters Institute Digital News Report?

It's an annual survey tracking global news consumption patterns, now in its 15th year. It surveys tens of thousands of people across 40+ countries.

Source: Fast Company / Pete Pachal