As the amount of available data continues to grow at an exponential rate, it seems to be impractical for humans to manually process and review this information. Thus, recommender systems are aiding the process of decision-making and discovering new content. In today's digital landscape, recommender systems have become an integral part of the user experience on platforms such as Google, Netflix, Spotify and Amazon. This is why it is important that users have confidence in the suggestions made by these systems. To foster such trust, transparency in how recommendations are generated is essential. Therefore, users should not only understand why certain suggestions are made but also have the possibility to interact with the system to refine and argue about these recommendations. Interacting in such a way may help to build trust by challenging the system instead of blindly trusting it.

The aim of this Shonan Meeting is to explore how recommender systems can be enhanced through argumentation and explanation capabilities, and how these enhancements impact users' trust and conversational experience. We will begin by discussing current recommendation technologies, including the ways in which conversational systems facilitate users in achieving recommendation-related goals through multi-turn dialogue. Building on this foundation, we will examine contemporary approaches to argument design and explainable AI (XAI), and evaluate their effects on perceived trustworthiness, usability, and ethical considerations.

Without any doubt, there are several challenges in building trustworthy and interactive recommendation systems. The main challenge is to develop technological models, methods and strategies for designing trustworthy interactive recommendation systems using argumentation, thereby addressing several aspects:

Recommendations made by trustworthy and interactive recommender systems need to be made transparent using methods from the field of XAI. Furthermore, these explanations can be substantiated by including general information available through information retrieval and argument mining. The integration of state-of-the-art large language models further enhances the system's natural language understanding, generation and contextual reasoning capabilities. By implementing interaction based on conversation and natural language, users can intuitively communicate with the system. To this end, the dialogue and argumentation strategy applied should consider factors such as persuasion, negotiation, proactivity and awareness of social, cultural and contextual circumstances.

Information and argument retrieval represent another key challenge. The use of advanced information retrieval techniques and natural language processing makes it possible to construct persuasive arguments from large data sets. Information and argument retrieval are complemented by the need for knowledge reasoning, where sophisticated reasoning skills help to understand complex relationships and make accurate recommendations.

Finally, the role of multimodal input and output, including speech, gesture and emotion, cannot be overstated. Using different sensory data allows us to create more engaging and effective interactions.

An important discussion topic will be what impact these advanced technologies will have on society, end users, and industry. For industry, integrating argumentation and XAI in recommender systems potentially enhances trust, user engagement, and competitive differentiation by providing transparent, personalized, and context-aware recommendations. For society, these advancements could promote inclusivity, digital literacy, and ethical AI adoption, fostering trust in AI systems while addressing biases and accessibility. For end users, transparent and interactive recommendations may improve satisfaction by enabling users to refine suggestions, make informed decisions, and enjoy intuitive, multimodal experiences. However, we also need to discuss the potential negative impact on society and end users in particular. For example, for society, poorly designed systems may perpetuate biases, raise privacy concerns due to increased data collection, and widen the digital divide by favoring resource-rich users. For end users, such systems could lead to cognitive overload, frustration with complex interactions, or a loss of autonomy if argumentation feels manipulative.

We have decided to structure the workshop into four different areas:(1) Research Challenges, (2) Use cases, user groups and industrial applications,(3)Development, testing and evaluation, and (4) Ethics and societal impact:

● Explainability of recommendations
● Trust (neither under- nor over-trust the system)
● Persuasion and negotiation
● Proactivity
● Argument quality
● Personalization
● Social- and context-awareness
● Cross-domain applicability
● Information retrieval and argument mining
● Knowledge Reasoning
● Informed decision making
● Interaction based on conversation and natural language
● Role of multi-modal in- and output (speech, gestures and emotions)
● How far can we go with state-of-the-art large language models

●Appropriate user groups (e.g. elderly, youngsters, and school kids)
●Specific application domains
　　○public space (e.g., interactive digital signage, guidance systems)
　　○assistive environments (e.g. elderly care, hospitals)
　　○education and tutoring
　　○computer-mediated human-to-human interaction
　　○navigation systems
　　○decision support systems
●Success stories, functional systems and industrial challenges
●Accounting for cultural differences

● Experimental design, user studies and evaluation
● Investigation of long-term vs. short-term relations
● Novel evaluation paradigms

● Social responsibility
● Legal issues
● Data protection and privacy by design and default
● Social design and development of naturally interacting human-recommender systems

Building trustworthy and interactive recommender systems requires an active collaboration between multiple research disciplines, such as computer science, psychology, and ethics.

This will result in a platform to exchange ideas and benefit from complementary work. We aim to create a unique venue for discussion and collaboration between experts from these disciplines. This Shonan Meeting will help to explore possible challenges and jointly develop a research agenda for main directions. Therefore, we will invite keynote speakers from the respective research fields, whose contributions will serve as a basis for breakout sessions.

In these sessions, participants will work actively on specific research objectives in small groups. This will help foster an interdisciplinary understanding and cooperativity.

The results of the breakout sessions will then be discussed with the whole plenum. The outcome of the workshop will be published in the form of free open-access CEUR-Proceedings (http://ceur-ws.org/). This is expected to encourage joint publications at top conferences and journals in computer science, jointly authored by psychologists, ethicists, and AI researchers. The planned Shonan Meeting will help to explore possible challenges and jointly develop aresearch agenda for main directions. The meeting will establish a platform for an international collaboration for the next three to five years.