Harnessing the power of AI innovation to inspire health communication messages that elevate awareness of the cancer risks linked to alcohol.

367 Background: Alcohol consumption is the third leading cause of cancer in the US and is classified as a Group 1 carcinogen linked to seven cancers, including mouth, throat, larynx, liver, colon, rectum, and breast cancers. Only 30% of Americans are aware of this risk. This study aims to develop and test AI-generated pictorial messages to raise awareness about alcohol-related cancers. Methods: We applied AI tools to generate the messages. Adobe Firefly and CoPilot were used for generations, with final touch-ups in Adobe Photoshop. Given the social acceptance of alcohol, messages were developed in 3 intensities: text message (TM), neutral picture message (NM), and graphic picture message (GM) (8x3 = 24 messages). We conducted an online factorial experiment with current alcohol users (n = 640; 49.3% female; age 21 + ; consumed ≥1 drink in the past month) to evaluate the GMs and NMs vs. TMs on health communication outcomes (attention, avoidance, harm perception, intention to reduce or stop drinking). Each participant viewed 2 of 8 messages, each shown at 3 intensities, totaling 6. We used the Friedman test for planned comparisons, followed by Wilcoxon signed-rank tests for multiple comparisons. Results: The average AUDIT-C score was higher among males than females (5.97 vs. 4.82). Heavy and younger drinkers were more likely to possess positive social norms and beliefs regarding the health benefits of alcohol. While 62.3% recognized the link between alcohol and liver cancer, awareness of other related cancers was low (26.6% colon cancer, 26.6% laryngeal cancer, and 20.8% breast cancer). Although GM about colon and throat cancers received the highest rating for avoidance, it received the highest rating for raising awareness, increasing intention to reduce or stop alcohol consumption. GMs significantly outperformed TMs in all outcomes, whereas NMs did not in terms of intention to change behavior (i.e., reducing, stopping, or committing to limited drinking goals). Conclusions: AI-generated messages can effectively raise awareness about alcohol-related cancer risks and promote healthier drinking habits. This research introduces a novel approach to health communication, focusing on cancer prevention, and utilizes advanced generative AI tools to create messages for population health campaigns, policy development, and provider-patient counseling.