When Machines Speak Science: Testing Consumers’ Perceptions of AI-Generated Communication Messages

Alexa J. Lamm, University of GeorgiaFollow
Kevan W. Lamm, University of GeorgiaFollow
Allison R. Byrd, University of GeorgiaFollow
Masoud Yazdanpanah, University of GeorgiaFollow
Nicholas K. Gabler, Iowa State UniversityFollow
Anna K. Johnson, Iowa State UniversityFollow
Catherine Sanders, North Carolina State University at RaleighFollow
Fallys Masambuka-Kanchewa, Iowa State UniversityFollow
Michael S. Retallick, Iowa State UniversityFollow

Abstract

Effective science communication is essential for fostering public understanding and trust in scientific information, particularly as emerging technologies like artificial intelligence (AI) are used to generate content quickly. The purpose of this research was to explore how U.S. consumers perceive transparency and source credibility of an agricultural science communication message delivered via Facebook when messages are generated by AI versus a human scientist. Using an experimental design embedded in an online survey, 1,011 respondents, representative of the U.S. public, were randomly assigned to view one of four social media posts covering either a benign (animal nutrition) or controversial (animal welfare) topic, authored by AI or a scientist. Respondents rated the perceived transparency of the messages before and after the source was revealed. Results indicated perceived transparency significantly decreased when respondents discovered the post was AI-generated and increased when they learned it was scientist-generated. Topic contentiousness did not have a significant influence. Findings suggest a preference for human-generated science communication messages, emphasizing the importance of source disclosure and highlighting challenges to public acceptance of AI-generated messaging. Science communicators should be cautious using AI when communicating science to the public until trust and acceptance of AI improves.

Recommended Citation

Lamm, Alexa J.; Lamm, Kevan W.; Byrd, Allison R.; Yazdanpanah, Masoud; Gabler, Nicholas K.; Johnson, Anna K.; Sanders, Catherine; Masambuka-Kanchewa, Fallys; and Retallick, Michael S. (2025) "When Machines Speak Science: Testing Consumers’ Perceptions of AI-Generated Communication Messages," Journal of Applied Communications: Vol. 109: Iss. 3. https://doi.org/10.4148/1051-0834.2594