A new study,  Are all uses of AI created equal? An experimental review of AI disclosure types on credibility, that examines how different AI disclosure labels influence audience perceptions and whether the wording of those disclosures affects credibility, was published in Journalism.  

That work was conducted by School Director and Professor, Dr. Patrick Merle, alongside graduate student Mitch Krueger and Assistant Professor of Public Relations at Syracuse University, Dr. Erika J. Schneider. 

“This project started from discussions with some of my master’s students when we asked ourselves how news organizations may prefer to let their audiences know about their AI usage,” said Merle. 

The study found that people viewed messages labeled as AI-generated as less credible than those described as AI-assisted or AI-influenced. Researchers were surprised to find that AI literacy had a smaller impact on credibility judgments than expected. 

The findings highlight how the language used to disclose AI involvement can shape audience perceptions. As AI tools become increasingly common in journalism and communication, understanding how these disclosures influence trust is becoming more important for media professionals. 

The authors hope to encourage news professionals to continue to question how to transparently communicate their AI usage to audiences. Because audiences respond differently to different types of AI involvement, there is no one-size-fits-all approach to disclosure, making it important to understand where such nuances exist and how to tell that to audiences.  

As AI continues to evolve, Merle hopes the study will serve as a resource for journalists, communicators and researchers navigating the role of transparency in the field.