The use of artificial intelligence in the fight against tobacco on social networks

In a commentary published in Tobacco Control[1], researchers from Yale School of Medicine and the University of Texas at Austin are exploring the potential of generative artificial intelligence (AI) as a tool to combat smoking on social media platforms. But they caution against its overuse, highlighting the fine line between its benefits and risks.

The study highlights the rapid growth of tools such as ChatGPT, which now has more than 180 million users, and their potential to address public health challenges. “Generative AI is more than a technological novelty; it is a disruptor that has the power to shape societal behaviors,” notes Dr. Grace Kong, the study's lead author.

The ubiquitous use of social media by adolescents and young adults, the significant presence of tobacco and vaping-related content, and the large body of data linking exposure to such content to an increased likelihood of nicotine initiation make social media an important target for tobacco surveillance. Similarly, they provide a platform for the dissemination of tobacco prevention interventions and messages, according to the authors.

Powerful tools to monitor tobacco advertising and provide quit tips

For several years now, social media has been used by the tobacco industry to disseminate numerous contents in favor of tobacco and other nicotine products.

The commentary highlights the potential of generative AI to revolutionize how tobacco-related content is monitored and addressed. AI-powered tools can analyze large amounts of social media data, identifying trends, misinformation, and marketing strategies quickly and efficiently. Currently, tobacco and nicotine-related content and reactions to it on social media are typically analyzed using qualitative content analysis conducted by humans. AI tools thus offer significant potential for researchers. They can be applied to public health and tobacco control by using generative AI to rapidly filter and analyze content for timely monitoring and development of counter-messaging to these same targets.

Generative AI tools could also improve interventions, for example by creating personalized chatbots for smoking cessation. “Chatbots integrated into social media platforms can provide targeted and timely responses, connecting users with resources and encouraging smoking cessation,” suggest the authors. However, understanding the interactions between users and AI-generated content remains crucial to refining these interventions.

Ethical challenges and concerns

Despite its promise, generative AI poses significant challenges, the authors say. One major concern is the technology’s “double jeopardy”: While it can counter tobacco promotion on social media, it can also amplify pro-smoking/vaping messages on those same platforms. The tobacco industry also leverages AI to spread misinformation or create large-scale marketing campaigns. Additionally, AI companies often fail to disclose the data sources they use or how they are exploiting it, raising concerns about privacy and user exploitation.

The commentary highlights the need for transparent evaluation frameworks to ensure that AI results are reliable, unbiased, and ethical, but also for human oversight to preserve the scientific integrity and validity of the results. Policymakers and researchers must work together to develop guidelines that prioritize public health while keeping pace with technological advances, the authors say.

©Tobacco Free Generation

AE

[1] Kong G, Ouellette RR, Murthy D. Tob Control Epub ahead of print: doi:10.1136/ tc-2024-058813