Can AI Improve Public Awareness of Alcohol-Related Cancer Risks?

By: Staff Writer | August 27, 2026 | 7 min. read |

A study from Sylvester Comprehensive Cancer Center found that cancer-specific messages paired with carefully reviewed, AI-generated visuals attracted more attention and increased perceptions of alcohol-related harm.

Three public-health communication cards displayed side-by-side on a contemporary review surface, illustrating different warning-message design approaches. The left card uses a simple text-focused layout, the center card incorporates visual storytelling elements and icons, and the right card features a more visually striking design with high contrast and complex imagery. In the background, blurred analytical dashboards, comparison graphics, and abstract network patterns suggest evaluation of message effectiveness. Subtle orange and green accents appear throughout the composition. The image conveys side-by-side assessment of communication strategies in a research and AI-assisted design context.

Most Americans understand that smoking causes cancer. Far fewer know alcohol is linked to several cancers.

A study from Sylvester Comprehensive Cancer Center, part of the University of Miami Miller School of Medicine, found that cancer-specific messages paired with AI-generated visuals may help people better understand alcohol-related cancer risks.

Alcohol is classified as a Group 1 carcinogen and is associated with breast, colorectal, liver, mouth, throat, laryngeal and esophageal cancers. Yet awareness remains low, especially for breast, colorectal and throat cancers. Only about one in five study participants identified alcohol as a breast cancer risk factor.

“Many people remain unaware that alcohol is linked to several cancers,” said Sylvester researcher Taghrid Asfar, M.D., MSPH, lead author of the study and professor of public health sciences at the Miller School. “Our findings suggest that clear, evidence-based communication, particularly when it includes specific information about cancer risk, may be a useful way to improve public understanding of this important health issue.”

Key Takeaways

Messages that named specific cancers generally performed better than broad cancer warnings.
AI-generated visuals outperformed text-only messages across most measures.
Graphic visuals produced the strongest reactions and heightened perceptions of alcohol-related harm.
The study measured engagement and intentions, not long-term knowledge or changes in drinking behavior.

How the Study Worked

Researchers from both Sylvester and the Miller School surveyed 639 U.S. adults age 21 and older who reported drinking alcohol. Participants saw messages about alcohol and cancer, then responded to questions about attention, engagement, perceived harm and intentions after viewing the materials. The study, which is published in Cancer, a journal of the American Cancer Society, did not measure actual changes in drinking behavior.

Some messages were general and others named specific cancers, including breast, colon, liver and throat cancer. Researchers tested text-only messages, messages with neutral AI-generated images and messages with more graphic AI-generated visuals.

What Worked Best

Messages that named particular cancers generally performed better than broad warnings about cancer risk. Participants were more likely to describe cancer-specific messages as informative, believable and attention-grabbing, especially when the messages included visuals.

Dr. Taghrid Asfar in dark jacket and white blouse — Dr. Taghrid Asfar says clear messaging can improve people’s knowledge about the links between alcohol and cancer.

Images consistently outperformed text-only messages across most measures. Graphic visuals generated the strongest reactions and heightened perceptions of alcohol-related harm. The study shows which messages people found more engaging. It does not show whether those messages changed long-term knowledge or drinking behavior.

Frequently Asked Questions

Is alcohol associated with cancer?

Alcohol is classified as a Group 1 carcinogen and is associated with breast, colorectal, liver, mouth, throat, laryngeal and esophageal cancers.

What messages worked best in the study?

Messages that named specific cancers generally performed better than broad cancer warnings, especially when they included visuals.

How was artificial intelligence used?

Researchers used AI tools to create visual concepts and message variations. Researchers and communications professionals reviewed and refined the images.

Did the messages change drinking behavior?

The study did not measure actual changes in drinking behavior or long-term knowledge. It measured responses such as attention, engagement, perceived harm and intentions.

Reaching Different Audiences

Moderate alcohol consumers responded more strongly than heavy consumers. Heavy drinkers were less likely to engage with the messages or say they intended to change behavior. Researchers say future work should explore how to communicate risk to different audiences without stigma or judgment.

Alcohol can be difficult to address in public health messaging because it is widely accepted in social and cultural settings.

Infographic titled “What Made Alcohol and Cancer Messages More Engaging?” presented in a four-step horizontal flow using University of Miami-inspired green and orange colors. Panel 1 compares a general cancer warning message with a more prominent cancer-specific warning message, showing that cancer-specific content is more impactful. Panel 2 illustrates three presentation formats: text-only, message with a neutral AI-generated visual, and message with a graphic AI-generated visual. Panel 3 displays participant responses across four measures: attention, engagement, believability, and perceived harm, with response levels increasing from text-only to neutral visuals to graphic visuals. Panel 4 summarizes study findings: cancer-specific messages performed better than broad warnings, images outperformed text-only messages, graphic visuals generated the strongest reactions, and moderate drinkers responded more strongly than heavy drinkers. A highlighted note at the bottom states that the study measured engagement, attention, and perceptions but did not assess long-term knowledge retention or changes in drinking behavior. The design uses icons, charts, arrows, and callout boxes in a clean academic medicine infographic style.

“Effective cancer prevention communication must be tailored to where people are psychologically, their readiness to adopt behavioral changes and their understanding of the relative risks involved,” said co-author Frank Penedo, Ph.D., director of the Sylvester Survivorship and Supportive Care Institute, associate director for population sciences, the Sylvester DCC Living Proof Endowed Chair in Cancer Survivorship and an associate professor in the University of Miami Department of Psychology. “This study underscores the importance of testing messages so we can better understand what captures attention, what builds trust and what may help people make informed decisions about their health.”

How AI Helped

Researchers used artificial intelligence tools to create visual concepts and message variations for testing. Researchers and communications professionals reviewed and refined the images to ensure accuracy, clarity and appropriateness for public health education.

More research is needed to determine whether these approaches improve knowledge retention or influence behavior over time.

Clear, specific messages supported by carefully reviewed visuals may help public health professionals communicate alcohol-related cancer risks more effectively.

“Our goal is not simply to tell people that alcohol can increase cancer risk, but to understand how to communicate that information in ways people notice, trust and remember,” Dr. Asfar said. “As we continue this work, we hope these findings can help shape public health messages that are clear, evidence-based and useful for people making decisions about their health.”