Generative AI chatbot vs telephone counseling for youth tobacco cessation: a pragmatic pilot randomized trial
This pragmatic pilot randomized trial demonstrated that implementing a generative AI chatbot for youth tobacco cessation within a quitline setting is feasible and yielded comparable biochemically validated abstinence rates and satisfaction scores to standard peer-led telephone counseling.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Most people who use tobacco start before they turn twenty-five. This early start makes it harder to quit later in life and increases the risk of dying young. While doctors have found effective ways to help adults stop smoking, those same methods often do not work well for young people. Young smokers behave differently and face different challenges than adults, so researchers have been searching for new ways to help them. In recent years, computer programs that can talk to people like humans have become very advanced. These programs, known as generative artificial intelligence, can understand what a person types and write back with original, helpful responses rather than just reading from a fixed script. This technology offers a way to provide support to many young people at once, but no one knew if it could actually help them stop using tobacco.
To find out, a team of researchers in Hong Kong set up a real-world test involving young people who wanted to quit. They recruited 154 participants, all under the age of twenty-five, who were currently using tobacco products like cigarettes or electronic cigarettes. The researchers split these young people into two groups. One group received the standard help available in their city: a series of phone calls with trained peer counselors who used proven talking techniques to encourage quitting. The other group used a new tool: a web-based chatbot powered by advanced artificial intelligence. This chatbot was designed to act like a counselor, asking questions, listening to answers, and offering advice on how to quit, all through text messages on a smartphone or computer. The chatbot was built to follow the same professional guidelines as the human counselors, but it could talk to anyone at any time without needing to wait for a person to be available.
The study tracked these young people for six months to see who managed to stop using tobacco. To be sure the results were real, the researchers did not just ask the participants if they had quit; they also tested their saliva for a chemical that appears in the body when someone uses tobacco. This chemical test confirmed that the young people who said they had quit were telling the truth. The results showed that the chatbot was a feasible tool to use. It was easy to get young people to sign up for the study, and most of them stayed in the study until the end. In the group using the chatbot, 71 percent of the participants used it at least once, and 62 percent completed their first counseling session. The young people who used the chatbot found it easy to use and rated its usability as satisfactory.
When it came to the actual goal of stopping tobacco use, the numbers suggested the chatbot might be just as effective as the human counselors, though the study was too small to say this with absolute certainty. At the six-month mark, 23 percent of the young people in the chatbot group had passed the chemical test for tobacco abstinence, compared to 17 percent in the group that spoke with humans. While the chatbot group had a higher success rate, the difference was not large enough to prove that one method was definitively better than the other. The researchers noted that the chatbot group did not complete as many sessions as the phone group, because the chatbot required the young people to log in and type on their own, whereas the phone calls were made directly to them by a counselor. Despite this difference in how they engaged with the help, the young people in both groups reported feeling equally satisfied with the support they received.
This trial was the first time such a test had been done with this specific type of advanced artificial intelligence for young smokers. The researchers found that it is possible to run this kind of program within an existing service for youth. The chatbot did not produce any dangerous or wildly incorrect answers during the study, which is a major safety concern when using artificial intelligence. However, because the study included a relatively small number of people, the results are best seen as a promising sign rather than a final answer. The data suggests that a chatbot could be a useful addition to the tools available to help young people quit, but larger studies are needed to confirm whether it works better than, or just as well as, talking to a human counselor. For now, the study shows that this new technology can enter the conversation and offer a viable path for young people seeking to leave tobacco behind.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.