Motivational Interviewing (MI) is a counseling technique that promotes behavioral change through reflective responses to mirror or refine client statements. While advanced Large Language Models (LLMs) can generate engaging dialogues, challenges remain for applying them in a sensitive context such as MI. This work assesses the potential of LLMs to generate MI reflections via three LLMs: GPT-4, Llama-2, and BLOOM, and explores the effect of dialogue context size and integration of MI strategies for reflection generation by LLMs. We conduct evaluations using both automatic metrics and human judges on four criteria: appropriateness, relevance, engagement, and naturalness, to assess whether these LLMs can accurately generate the nuanced therapeutic communication required in MI. While we demonstrate LLMs’ potential in generating MI reflections comparable to human therapists, content analysis shows that significant challenges remain. By identifying the strengths and limitations of LLMs in generating empathetic and contextually appropriate reflections in MI, this work contributes to the ongoing dialogue in enhancing LLM’s role in therapeutic counseling.

People increasingly have conversational agents at home. These are often shaped as boxes or tablets, and used for single question/answer interactions. This paper describes a virtual agent designed as a friendly neighbor, shown in a painting frame, available when needed for extended dialogues including generative AI model support, and giving privacy at other times.

Chatbots are becoming omnipresent in our daily lives. Despite rapid improvements in natural language processing in the last years, the technology behind chatbots is still not completely mature, and chatbots still make a lot of mistakes during their interactions with users. Since it is not possible to completely prevent mistakes due to technological constraints, this article aims to investigate whether a human-like communication style can reduce the negative impact of chatbots’ mistakes on users. Taking a combination of the Technology Acceptance Model and the concepts of Perceived Enjoyment and Social Presence as a theoretical basis, we conducted an online experiment in which participants interacted with a chatbot and completed a survey afterwards. We found that chatbot mistakes have a negative effect on users’ perceptions of Ease of Use, Usefulness, Enjoyment, and Social Presence. Human-like communication was found to be effective in reducing the negative impact of mistakes on Perceived Enjoyment. Theoretical and practical implications are discussed.

Robots are becoming an integral part of society, and might soon take on roles involving making morally relevant decisions. In a pre-registered experiment (n = 184), we investigated which factors modulate the extent to which we trust a robot to make a moral choice. Specifically, the effects of anthropomorphic appearance and anthropomorphic agency and affect attributions were assessed. Participants were presented with moral dilemmas in which the individual having to make a decision was a humanoid or mechanical robot. Each robot was described in vignettes in which they were attributed with agency and/or affective states. Subsequently, participants’ implicit moral trust in the robot was measured, as well as explicit trust, perceived capability of the robot, and the extent to which they felt the robot was responsible for its choice. Both agency and affective state attributions were found to impact participants’ implicit and explicit trust as well as the perceived capability of the robot. Moreover, across conditions, mechanical robots were trusted significantly more than humanoid robots to take moral choices.

Contains fulltext : 250750.pdf (Publisher’s version ) (Open Access)

This article is part of a project that explores the potential of chatbots for providing online emotional support to humans tailored to stressors. Based on a number of empirical studies, we have developed a socially interactive agent able to have simple dialogues with stressed humans seeking for emotional support. In the current article, we address the question to what extent this chatbot is effective in helping users cope with stressful situations. To this end, we present a study in which participants were asked to interact with our proposed chatbot for three days. Participants are distributed over the following three conditions—namely: 1) receiving support from the chatbot, knowing the support is computer-generated; 2) receiving support from the chatbot, while believing the support is human-generated; 3) not receiving any support. During the three days, participants’ self-reported stress levels are measured on a daily basis before and after each interaction. Results indicate that the best results are obtained in the ‘human’ condition, while the worst results are obtained in the “computer” condition. These findings lead us to conclude that the presumed sender of a stress support message (i.e., a human or a computer) might be more important than the content of the message.

There is a new crisis emerging in human-agent interaction research: Instead of using validated questionnaires, individual studies are creating new questionnaires that claim to measure identical constructs. This makes replication studies and comparisons between studies near to impossible. In turn, meta-analyses to determine which characteristics are important to create agents that the user experiences as being intelligent are difficult to conduct. As part of the attempt to battle this crisis, in this current paper, we suggest the use of a Theory of Mind task to measure the implicit social behaviour users exhibit towards a virtual agent. In a two-part study, we present findings that suggest that participants conduct this Theory of Mind task as expected: participants adapt towards our virtual agent more than when they conduct the task alone. We additionally present preliminary results correlating performance in the Theory of Mind task to validated constructs. Unfortunately, our current results do not correlate significantly to the existing constructs. Data-collection is ongoing and hence no firm conclusions can be made about this second set of results. However, our data suggest that it is important to become aware that the existing validated constructs used in HCI research may not be tapping into what the researchers assume, and hence provide a basis for important discussions about these implications.

Artificial Intelligence techniques are increasingly being used to develop smart training applications for professionals in various domains. This paper presents an intelligent training system that enables professionals in the public domain to practice their aggression de-escalation skills. The system is one of the main products of the STRESS project, an interdisciplinary research project involving partners from academia, industry and society. The system makes use of a variety of AI-related techniques, including simulation, virtual agents, sensor fusion, model-based analysis and adaptive support. A preliminary evaluation of the system has been conducted with two groups of potential end users, namely tram conductors and police academy students.

To endow game characters with more realistic affective behavior, the notion of prospect-based emotions plays an important role: recent literature suggests that emotional states of such agents should not only be triggered by present stimuli, but also by anticipation on future stimuli, and evaluation of past stimuli in the context of these anticipations. Within the current study, an extension of the belief-desire-intention (BDI) model with prospect-based emotions is proposed, and is evaluated with respect to its capabilities of enhancing believability of game characters. The model has been implemented in the modeling language LEADSTO. In addition, a game application has been developed, in which a user can play a two games (tic-tac-toe and a game of dice) against an agent that is equipped with the emotion-based model. An empirical evaluation indicates that the model significantly enhances the agent's believability, in particular concerning its involvement in the situation.

Collective decision making involves on the one hand individual mental states such as beliefs, emotions and intentions, and on the other hand interaction with others with possibly different mental states. Achieving a satisfactory common group decision on which all agree requires that such mental states are adapted to each other by social interaction. Recent developments in social neuroscience have revealed neural mechanisms by which such mutual adaptation can be realised. These mechanisms not only enable intentions to converge to an emerging common decision, but at the same time enable to achieve shared underlying individual beliefs and emotions. This paper presents a computational model for such processes. As an application of the model, an agent-based analysis was made of patterns in crowd behaviour, in particular to simulate a real-life incident that took place on May 4, 2010 in Amsterdam. From available video material and witness reports, useful empirical data were extracted. Similar patterns were achieved in simulations, whereby some of the parameters of the model were tuned to the case addressed, and most parameters were assigned default values. The results show the inclusion of contagion of belief, emotion, and intention states of agents results in better reproduction of the incident than non-inclusion.

In this special section we report on the workshop Standards in Emotion Modeling held in August 2011, Leiden, The Netherlands. An important goal of this workshop was to identify challenges related to the development and evaluation of computational models of affective processes. These challenges were approached from a psychological and a computational modeling perspective. In this introduction we present a summary of the results of this week-long workshop. In addition to that, we are proud to present an invited contribution that proposes solutions to several of these challenges.

This paper discusses and evaluates an agent model that is able to manipulate the visual attention of a human, in order to support naval crew. The agent model consists of four sub-models, including a model to reason about a subject's attention. The mo

To enhance believability of virtual agents, this paper presents an agent-based modelling approach for decision making, which integrates rational reasoning based on means-end analysis with personal psychological and biological aspects. The agent model developed is a combination of a BDI-model and a utility-based decision model in the context of specific desires and beliefs. The approach is illustrated by addressing the behaviour of violent criminals, thereby creating a model for virtual criminals. Within a number of simulation experiments, the model has been tested in the context of a street robbery scenario. In addition, a user study has been performed, which confirms the fact that the model enhances believability of virtual agents.

Central research questions addressed within Criminology are how the geographical displacement of crime can be understood, explained, and predicted. The process of crime displacement is usually explained by referring to the interaction of three types

To endow virtual agents with more believable affective behavior, it is important to provide them not only the capability to generate and regulate their own emotions, but also the ability to reason about the emotion regulation processes of the agents

In recent times researchers have initiated investigating emotion as a collective property of groups, emphasizing the influence of combined emotions among group members on group processes. Within groups humans recognize and react emotionally to expressions of emotions of other group members. This paper uses a multi-agent-based approach to formalize and simulate such emotion contagion within groups. 1.

Within many domains, among which biological, cognitive, and social areas, multiple interacting processes occur among agents with dynamics that are hard to handle. This paper presents the predicate logical Temporal Trace Language (TTL) for the formal specification and analysis of dynamic properties of agents and multi-agent systems. This language supports the specification of both qualitative and quantitative aspects, and therefore subsumes specification languages based on differential equations and qualitative, logical approaches. A software environment has been developed for TTL, which supports editing TTL properties and enables the formal verification of properties against a set of traces. The TTL environment proved its value in a number of projects within different biological, cognitive and social domains.

In this paper a simulation model for visual attention is discussed and formally analysed. The model is part of the design of an agent-based system that supports a naval officer in its task to compile a tactical picture of the situation in the field.

Within Criminology, the process of crime displacement is usually explained by referring to the interaction of three types of agents: criminals, passers-by, and guardians. Most existing simulation models of this process are agent-based. However, when the number of agents considered becomes large, population-based simulation has computational advantages over agent-based simulation. This paper presents both an agent-based and a population-based simulation model of crime displacement, and reports a comparative evaluation of the two models. In addition, an approach is put forward to analyse the behaviour of both models by means of formal techniques.

When developing sophisticated multi-agent systems whose behaviors include collaboration, negotiation, and conflict resolution, analyzing and (empirically) verifying agent system behavior is a challenging task. To aid the developer in such tasks, this