Artificial intelligence is transforming how information is created, consumed, and trusted. As AI makes misinformation becomes to produce and harder to detect, researchers are beginning to uncover how these shifts affect news consumption, audience trust, and the role of established publishers. 

A new study from the National Bureau of Economic Research examines how AI-generated misinformation affects audience behavior and publisher strategies. Working with the major German news outlet Süddeutsche Zeitung (SZ), the report, GenAI Misinformation, Trust, and News Consumption: Evidence from a Field Experiment, provides timely evidence on the challenges and opportunities facing trusted media brands. 

The results reveal an important paradox. While exposure to AI-driven misinformation raises concerns about trust in media overall, it also increases engagement and strengthens loyalty to trusted news brands. The findings point to both challenges and opportunities in an era where credibility is increasingly scarce. 

Experiment context 

This extensive study includes more than 17,000 individual surveys measuring SZ reader concern with misinformation, trust in various media outlets, and willingness to pay for subscriptions. In addition, approximately 6,000 participants consented to have their browsing and subscription data tracked. And a subset of respondents participated in a quiz featuring side-by-side images, some real and some generated by AI, to identify which images are authentic. 

Double-edged impact 

The findings underscore the dual nature of AI-driven misinformation. On one hand, it erodes trust in media content overall. On the other hand, it highlights the relative value of credible news organizations that can help audiences navigate a confusing information environment. 

• Concerns about misinformation: Participants’ exposure to an AI image quiz increases their worries about false content. 

• Trust declines: Trust in news falls across all outlets, including SZ itself, showing that awareness of AI-generated content reduces confidence in the broader news environment. 

• Engagement increases: Daily visits to SZ’s digital content grow by 2.5% immediately after the research, with the effect lasting more than two weeks before tapering off. 

• Subscriber retention improves: Five months after the experiment, participants in the treatment group are 1.1% more likely to remain subscribed, roughly a one-third reduction in attrition compared to the control group. 

• Impact is greatest for vulnerable readers: Readers struggling the most to distinguish real images from AI images show the strongest engagement and retention — daily visits rising 4% in the first three to five days and a 3.5% increase in willingness to pay for a subscription. 

The study finds that transparency and education efforts help strengthen audience loyalty. Although trust scores decline overall, readers who struggled with the AI quiz do not lose confidence in SZ. Instead, they visit the site more often and show greater willingness to pay for its content. This suggests that when people recognize the challenges of navigating misinformation, they lean more heavily on reliable sources. 

The bigger picture 

As noted, the findings highlight how trusted journalism drives engagement and reduces churn, even amid growing skepticism. Credibility emerges as a powerful business asset, as audiences turn to dependable brands to make sense of an uncertain information environment. As AI-driven misinformation makes navigating the news increasingly difficult, credible publishers are well-positioned to guide audiences, delivering the clarity, reliability, and trust that readers seek.