AI Awareness and Tobacco Policy Messaging Among US Adults: Electronic Experimental Study

Area of Science:

• Public Health
• Health Communication
• Artificial Intelligence

Background:

• Tobacco use is a leading cause of preventable death in the US, disproportionately affecting underrepresented populations.
• Public policies and health messaging are crucial for improving tobacco-related health equity.
• Artificial intelligence (AI) offers potential for rapid creation of tailored public health messages, but its impact on public perception is under-researched.

Purpose of the Study:

• To investigate how knowledge of artificial intelligence (AI) use influences public perception of a tobacco control policy video.
• To assess the impact of AI narrator disclosure on perceived video effectiveness, credibility, and speaker ratings.

Main Methods:

• A national sample of 500 US adults viewed an AI-generated tobacco control policy video.
• Participants were randomly assigned to one of four conditions: AI narrator, human narrator, unknown narrator, or no information.
• Perceptions of the video and narrator were measured across conditions.

Main Results:

• No significant differences were found in perceived video rating, effectiveness, or credibility across conditions.
• The speaker rating was significantly higher when participants believed the narrator was human compared to AI or unknown.
• Positive attitudes toward AI were highest when narrator information was withheld, though not statistically significant.

Conclusions:

• AI's impact on speaker perception in health messages warrants further investigation, particularly regarding long-term effects and repeated exposures.
• Qualitative research may elucidate reasons for differences in speaker ratings.
• Public health practitioners should consider the ethical and cost-effectiveness of employing AI in health messaging.