BACKGROUND: In Brazil, the online sale of tobacco products is prohibited, as the marketing, advertising and sale of Electronic Smoking Devices (ESD). The artificial intelligence (AI) use may offer tools for detecting and combating this illegal activity. The present study aims to develop a diagnosis of the illegal sale of tobacco products and ESDs on the Brazilian Internet.
METHODS: A technology company was hired to develop a text and image tracking methodology with machine learning, using Google Search, Apify and SerpApi. The Due to Instagram's encryption the search was only for texts. After the quantitative survey, a sample of the content of websites and Instagram profiles were analyzed.
RESULTS: The search found tobacco sale at:

• 706 websites;
• 44 websites (cigarettes);
• 305 Instagram.

The websites selling ESD, can be classified as:
• exclusively sale, as traditional e-commerce and;
• websites that combine sales and promotion, often using visual appeals aimed at young audiences. Figure 1.
The websites offer delivery and various payment methods. The websites that sell only cigarettes adopt a traditional marketing approach, without appealing to young people. The Instagram profiles were characterized as ESD points of sale, with delivery and aiming at young people. The ESD’s promotion on social networks is similar to the websites. Figure 2
Websites provide links to social networks such as Instagram, TikTok, Facebook and WhatsApp, highlighting the commercial integration between various platforms, intensifying the promotion of tobacco products.
CONCLUSIONS: The study shows that social media and e-commerce platforms are vectors promoting tobacco products and ENDS and the illicit sale of tobacco products on the Internet, remains in Brazil violating the current regulations, requiring the strengthening of online monitoring. The AI may strengthen surveillance by public authorities and a breakthrough capability in the analysis of voluminous amounts of data on the Internet, overcoming the boundaries of traditional monitoring methods.