'Smoke detector' developed against cigarette and tobacco promotion on social media! It will now be blocked

Public Health Researcher Dr. Hüseyin Küçükali began research 5 years ago at Istanbul University-Cerrahpaşa (İÜC) on social media content that promotes the use of cigarettes and other tobacco products.

SUPPORTED BY YEŞİLAY

Within the scope of the research, which is also supported by the Turkish Green Crescent (Yeşilay), AI-powered software was developed to detect content on social media that encourages tobacco use.

Dr. Hüseyin Küçükali stated that tobacco companies are trying to replace the hundreds of thousands of customers who die each year in order to maintain their profitability.

Stating that tobacco companies are therefore turning toward young people, Küçükali noted that there has been an acceleration in the promotion of tobacco products, especially with the electronic cigarettes that have emerged in recent years.

Pointing out that tobacco companies actively use social media to reach young people, Küçükali continued as follows:

"Unlike traditional media, social media carries certain difficulties in terms of detecting and preventing harmful content. In Turkey and around the world, the promotion of cigarettes and other tobacco products is blocked by laws regulating traditional media. Social media is difficult to control in this regard. We see that tobacco companies continue to promote tobacco products by using people called 'influencers' and guerrilla marketing techniques. In this study, we developed an artificial intelligence model that can automatically detect content on social media that glamorizes tobacco products and encourages their use."

THEY WILL NOW BE MONITORED ON SOCIAL MEDIA

Public Health Researcher Dr. Küçükali explained that they started the study in 2020 and are using "Large Language Models", which are widely used in tools like ChatGPT today.

Expressing that content reaches the masses on social media in a very short time, Küçükali said, "Even on older platforms like Facebook, content reaches the majority of the target audience within 1 hour. On newer social media platforms, it reaches the target audience in under an hour, in 10 minutes. This software, which automatically detects content on social media, will primarily make it possible to monitor this content. In other words, we will be able to report how much harmful content is on social media on which date, and we will be able to monitor changes over time."

Emphasizing that tobacco promotion is banned on social media in Turkey, Küçükali said they could detect content with the software and report it to the relevant authorities.

Küçükali pointed out that the AI-powered software they developed is essentially a "smoke detector" for social media.

Stating that the software will create an infrastructure upon which various applications can be developed by different institutions, Küçükali noted that it would be possible to integrate the AI software they developed into social media platforms and add a warning next to content that promotes tobacco.

USERS WILL BE ABLE TO MANAGE THEIR PREFERENCES

Küçükali provided the following information: "It could also be possible to use this software to empower the user. For example, we can talk about a smoke-free social media. A user who says 'I do not want to see content related to tobacco' can choose not to see it in the settings and block this content entirely."

Stating that he is continuing his research at Utrecht University, Küçükali added that their work is ongoing to expand the software to other commercial products harmful to health, such as alcohol, gambling, sugary drinks, and processed foods.