Behavior change is a complex research field, as modifying habits is difficult and often people relapse into their previous routines. Many current interactive applications appear to fail in providing long lasting change in behavior due to ineffective forms of “persuasion” that are not able to sustain the behavior change effort over time. On the one hand, behavior change requires consideration of elements that may differ across individuals, such as physical and mental states, skills, goals, risk factors, preferences, as well as of the fact that behavior change interventions have to be adjusted over time as the person progresses through the process of change in line with stages of change theory: in this sense, recommender systems exploiting the user’s personal data could provide tailored suggestions that may increase the likelihood of the effectiveness of the intervention.
On the other hand, however, the complexity of a behavior change recommender system is greater than that of other kinds of recommenders, due to the need of considering diverse contextual factors and incorporating multiple components for tracking, interacting, and personalizing: a pervasive behavior change recommender system monitors user behavior, such as mood and physical activity, considers user preferences to tailor recommendations, and prompts users and keeps them engaged to encourage adherence to interventions.
In this context, many theoretical and practical challenges and open issues arise: for instance, it is still not clear what kind of theories should be used to ground the design of behavior recommendations, what kind of interfaces and communication channels are more effective in delivering the recommendations, how the system can sustain the person’s motivation to initiate the change and adhere to the behavioral program, in which contexts (and at which times) the recommendation should or should not be delivered, and so on. Moreover, persuasive use of personalization and recommendations is causing increasing ethical and safety concerns about behavior engineering, which may harm human autonomy and wellbeing.
Therefore, in addition to the many opportunities that the current technological landscape provides for the design of novel recommender systems aiming at changing people’s behavior, it becomes urgent to discuss the different challenges that behavior change recommender systems should face in the near future. Despite the recent advances, there are many crucial research challenges ready for innovation. The workshop aims to provide a forum for discussing open problems and innovative research approaches in this area.
Some questions that motivate this workshop are: What kind of theory is more suitable to inform the design of behavior change recommender systems? What kind of data should be used to design recommendations? How should they be delivered? What kind of strategies should be implemented to design timely and contextualized recommendations? How to support the user’s motivation and help them sustain the desired behavior in the long term? What contextual factors may affect the effectiveness of recommender systems and should be considered in design?
behavrec.wordpress.com 