recommender systems. This volume includes the papers presented at BehavRec '23 International Workshop held in Singapore in conjunction with ACM RecSys '23 International Conference. The workshop aimed to discuss open problems, challenges, and innovative research approaches in the area of persuasive and behavior change Proceedings of BehavRec '23, the First International Workshop on Behavior Change and Persuasive Recommender Systems at the 17th ACM Conference on Recommender Systems (RecSys '23), Singapore, September 18-22, 2023 ∗Corresponding author.
Behavior change systems, also known as persuasive technologies, are interactive tools
intentionally designed to encourage people to modify their own behaviors and habits. They can
be used in a variety of domains, such as health [
According to a recent claim by IBM, 90 percent of the data available today have been created
in the last two years. This exponential growth of digital information has given new life to
research on persuasive technologies, allowing designers to deliver extremely personalized
behavioral interventions, potentially based on a variety of users’ psychological, physiological,
and behavioral data [
It therefore comes as no surprise that the interest in behavior change recommender systems,
which provide users with personalized recommendations on how to modify their behavior, has
considerably increased in recent years [
However, behavior change is a complex research field, as modifying behavior is extremely dificult and often the intervention fails for lack of engagement or simply because people relapse into previous habits. Moreover, behavior change requires consideration of extremely idiosyncratic factors that are peculiar for a specific individual, such as her objectives, motivations, risk factors, and preferences [16]: in this sense, recommender systems exploiting the user’s personal data could provide tailored suggestions that may increase the likelihood of the efectiveness of the intervention [17]. On the other hand, behavior change recommender systems are more complex than other kinds of recommenders, since they have to consider multiple contextual factors and incorporate components for monitoring user behavior, considering her preferences to tailor recommendations, and keeping her engaged to encourage adherence to the intervention [17].
For this, many new challenges and open issues arise: for example, we still do not know
what kind of theories we should use to ground the design of persuasive recommendations
[
Some questions that still need to be answered are: What kind of personal data should be used to design recommendations? How such recommendations should be delivered? What kind of theory is more suitable to inform their design? How can we support the users’ motivation and help them sustain the desired behavior in the long term? What contextual factors may afect the efectiveness of the recommendations and should be considered in design?
To this aim, topics of interests that need for further exploration are: 1) Diferent types of behavior change recommender systems and their peculiarities, like for health, wellness, safety, sustainability, etc. 2) User Interfaces for behavior change recommender systems, like visual interfaces, context-aware interfaces, ubiquitous, wearable and mobile interfaces, as well as conversational interfaces 3) New approaches to designing and delivering behavior change recommendations: Controllability, transparency, and explainability, argumentationaware recommendation, culture-aware recommendation, context-aware recommendation 4) How to balance the cost and benefit of behavior change recommender systems, as well as challenges and limitations of implementing them 4) Ethics, Privacy, and theories: Frameworks and models for developing personalized persuasive technology, as well as models for developing ethical and privacy-sensitive behavior change recommender systems - 5) Evaluation: Long-term evaluation and evidence of long-term efects of behavior change recommender systems.
The workshop addressed the aforementioned topics of interest. The organizing committee received eight papers, and six papers were accepted into the workshop proceedings. A short summary of each contribution is described below.
In “Towards an adaptive Behavior Change Game based on user-tailored and context-aware interventions”, Villata et al. propose to go beyond the one-size-fits-all approach often adopted in persuasive technologies, as well as to consider the meanings that users may attribute to the process of behavior change in the design process. They present a mobile-based adaptive Behavior Change Game that promotes positive habits by personalizing its gameplay on the basis of a comprehensive user model, containing information about the user’s habits, behavior and context collected through smart devices. The adaptation is performed for both the game content and interface, mainly changing, hiding or showing specific game design elements.
In “FitNExT: Leveraging Transformers with Context-Augmented Start Tokens to Generate Recommendations for New Users in Connected Fitness at Peloton”, Yoshida and Meetei introduce the Fitness New user Experience Transformer (FitNExT) model, which combines the strength of transformer architecture in understanding sequential data with an innovative approach for contextualizing the start of a member’s fitness journey. Then, they use the outputs of this model to display rows of recommendations to ease new users into their fitness routine.
In “Contextual Bandits for Hyper-Personalization based on User Behavior in Local Domain”, Kim et al. propose an empirical hyper-personalization problem reflecting user behavior in local domain, that is considered as contextual bandit problem with well-configured recommender system ensemble. The authors empirically introduce how to deal with insuficient user feedback in service by using feedback of other interfaces, how to define user contexts and user features in local domain, and how to ensemble contextual bandits for optimization.
In “Towards Adaptive and Personalised Recommendation for Healthy Food Promotion”, Nurbakova et al. present an adaptive persuasive system that exploits and extends the idea of a constrained question answering (QA) system over a knowledge graph proposed by Chen et al. They introduce the way to model personalised challenges and additional constraints allowing to handle meal plans.
In “Understanding How News Recommender Systems Influence Selective Exposure”, Seddik et al. attempt to ask the following research question: To what extent can News recommender systems influence the selective exposure behavior of news users? The authors present a preregistered online experiment to empirically test the impact of structural factors on selective exposure, by tracking users’ behavior on a news website equipped with two diferent versions of custom-made NRSs that are designed to nudge users towards increased or decreased selective exposure to like-minded or cross-cutting news.
Finally, in “Using persuasive strategies inside app distribution platforms to warn users about manipulative design used by applications” Babaei and Vassileva investigate the impact of employing persuasive strategies in mobile app distribution platforms to inform users about the manipulative design used by mobile applications. They present the design of a study with three groups, where they aim to measure the attitudes and intentions of the participants toward manipulative design using a questionnaire before and after the participants have been presented with an intervention. [12] M. Jesse, D. Jannach, Digital nudging with recommender systems: Survey and future directions, Computers in Human Behavior Reports 3 (2021) 100052. [13] A. Berge, V. V. Sjøen, A. Starke, C. Trattner, Changing salty food preferences with visual and textual explanations in a search interface, in: Proceedings of the ACM IUI 2021 Workshops, 2021. [14] C. Musto, C. Trattner, A. Starke, G. Semeraro, Towards a knowledge-aware food recommender system exploiting holistic user models, in: Proceedings of the 28th ACM conference on user modeling, adaptation and personalization, 2020, pp. 333–337. [15] A. Starke, C. Trattner, H. Bakken, M. Johannessen, V. Solberg, The cholesterol factor: Balancing accuracy and health in recipe recommendation through a nutrient-specific metric, in: Proceedings of the 1st Workshop on Multi-Objective Recommender Systems (MORS 2021), 2021. [16] A. Rapp, A. Boldi, Exploring the lived experience of behavior change technologies: Towards an existential model of behavior change for hci, ACM Trans. Comput.-Hum. Interact. 30 (2023). URL: https://doi.org/10.1145/3603497. doi:10.1145/3603497. [17] H. Torkamaan, J. Ziegler, Integrating behavior change and persuasive design theories into an example mobile health recommender system, Association for Computing Machinery, New York, NY, USA, 2021. URL: https://doi.org/10.1145/3460418.3479330. doi:10.1145/ 3460418.3479330. [18] A. El Majjodi, A. D. Starke, C. Trattner, Nudging towards health? examining the merits of nutrition labels and personalization in a recipe recommender system, in: Proceedings of the 30th ACM Conference on User Modeling, Adaptation and Personalization, 2022, pp. 48–56. [19] A. Starke, A. El Majjodi, C. Trattner, Boosting health? examining the role of nutrition labels and preference elicitation methods in food recommendation, in: Interfaces and Human Decision Making for Recommender Systems 2022, 2022, pp. 67–84. [20] A. Rapp, M. Tirassa, L. Tirabeni, Rethinking technologies for behavior change: A view from the inside of human change, ACM Transactions on Computer-Human Interaction (TOCHI) 26 (2019) 1–30. [21] S. Purpura, V. Schwanda, K. Williams, W. Stubler, P. Sengers, Fit4life: The design of a persuasive technology promoting healthy behavior and ideal weight, in: Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, CHI ’11, Association for Computing Machinery, New York, NY, USA, 2011, p. 423–432. URL: https://doi.org/10. 1145/1978942.1979003. doi:10.1145/1978942.1979003. [22] A. Rapp, Design fictions for behaviour change: exploring the long-term impacts of technology through the creation of fictional future prototypes, Behaviour & Information Technology 38 (2019) 244–272.