0
0
DELFT, THE NETHERLANDS – Researchers at Delft University of Technology have developed a smart AI coach that leverages psychologically informed reinforcement learning (RL) to help smokers and vapers quit. This innovative approach, developed by Nele Albers for her doctoral dissertation, utilizes machine learning to provide personalized support based on individual needs and behaviors.
Albers’ AI coach uses RL, a technique where the model learns through rewards, mirroring human learning processes. The development of this AI coach was informed by behavior change theories and data from three large-scale studies involving over 500 participants each.
Beyond the technical aspects, Albers’ research explored ethical, economic, and psychological factors that contribute to effective behavior change. She examined how to motivate smokers to quit, what interventions are most effective, and the role of support systems.
“My research shows how personalizing support—by taking into account both a person’s current and future state—increases the effectiveness of AI-based eHealth applications. This offers many opportunities for behavioral change,” Albers stated.
The AI coach aims to address challenges associated with traditional eHealth applications, such as high dropout rates and low engagement, by providing tailored support that considers individual knowledge, motivation, vitality, and thinking patterns.
A key aspect of Albers’ model is its ability to determine when human intervention is necessary, complementing the AI coach’s support. The algorithm also balances expert health advice with the smoker’s preferences, ensuring a more personalized and effective approach.
The research highlights the importance of adapting support to an individual’s situation. For example, the AI coach can suggest activities that address specific triggers or focus on future goals. Contextual messaging also plays a crucial role, with the AI coach adapting its language based on the situation. For instance, using phrases like “Other smokers find…” may be more effective than “Doctors recommend” in certain contexts.
By analyzing these factors, Albers developed models that enhance smoker engagement and provide better support for building skills to quit.
Disclaimer: This article is based on doctoral research from Delft University of Technology. While the AI coach shows promising results in helping smokers quit, it is important to understand that this is a research project, and further development and clinical trials are necessary before widespread implementation. The information provided in this article should not be interpreted as medical advice. Individuals seeking to quit smoking should consult with a healthcare professional for personalized guidance and support.
