Children increasingly interact with conversational AI in everyday settings, yet existing evaluation approaches rarely capture how trust, doubt, and understanding develop across childhood. This second edition of the C3AI workshop brings together researchers, designers, and educators to explore where trust, design, and evaluation meet in child-AI interaction. Through hands-on analysis of naturalistic interaction transcripts from children aged 6-14 and research-informed personas, participants will collaboratively identify cues of trust, confusion, resistance, and critique. Small-group activities and co-design exercises will support the development of preliminary, developmentally sensitive evaluation metrics for trust. Workshop outcomes include shared personas, draft evaluation templates, and open materials to support child-centred, responsible AI design and assessment.

Children increasingly encounter intelligent systems in educational and everyday contexts, yet most trust measures in human–AI interaction have been developed for adult users. This Work-in-Progress extends the validation of the child-centred K-AI Trust questionnaire to an older developmental cohort (10–14 years). Building on prior validation with 9–11 year-olds, we examine whether the instrument demonstrates acceptable internal consistency and coherent response patterns in early adolescence. Using survey data from 314 participants, we report preliminary reliability analyses and explore associations between dispositional trust (Propensity to Trust Technology; PTT) and post-interaction trust evaluations. Item refinement resulted in acceptable internal consistency for both scales (7-item K-AI; 5-item PTT). Correlational analyses revealed no significant association between dispositional and situational trust. This suggests that, in early adolescence, children’s inherent propensity to trust technology may not directly shape their immediate trust evaluation of a specific intelligent system. These findings provide preliminary evidence for the developmental applicability of a child-centred trust instrument and support ongoing age-sensitive psychometric refinement.

As generative AI systems increasingly act autonomously in digital environments, the ways in which these systems interact with children raise questions of both agency and permission. Existing frameworks offer limited guidance on how to design interfaces that respect children’s role as the final decision-makers [30]. This paper adopts a design-oriented perspective on child-AI interaction, drawing on mixed-methods data from 289 children aged 9-11. We reinterpret children’s trust-related responses as empirical signals for the ways in which agency and authority should be distributed within AI interfaces. Our findings indicate that children resist ‘proactive’ AI, instead preferring systems that seek permission, support non-judgmental error correction, and respect their right to refuse suggestions. Building on these insights, we propose three design principles: Staged Execution, Non-Destructive Reversibility, and Perceptible Uncertainty Signals, which translate children’s rights into concrete interface mechanisms, ensuring that generative systems support rather than override children’s agency.

Most trust metrics for intelligent systems are developed for adults, relying on complex reasoning and language that do not align with children’s developmental stages. As intelligent systems increasingly engage with young users, evaluating trust in child-AI interaction has become an urgent concern in HCI. In this paper, we present the iterative refinement and validation of the K-AI Trust Questionnaire, a child-centred instrument that integrates dispositional and situational trust components grounded in child-rights principles. Dispositional trust is captured through a child-adapted Propensity to Trust Technology (PTT), while situational trust is assessed through post-interaction items reflecting children’s experience with AI. Starting with a sample of 289 children, we conducted psychometric analyses and exploratory testing, culminating in a confirmatory factor analysis on a subsample of 85 children. Results supported a unidimensional structure consistent with the PTT, and highlighted the limitations of adult-oriented scales, underscoring the need for developmentally appropriate tools for trustworthy child-AI design.

This paper introduces OSMoSIS, a novel tool for assessing Social Motor Synchrony (SMS) in naturalistic child-facilitator interactions. Unlike Motion Energy Analysis (MEA), which is limited to seated or highly controlled tasks, OSMoSIS captures full-body interactions and computes synchrony through position correlation (PC) and specific kinetic energy correlation (SKEC). We evaluated OSMoSIS during music therapy sessions with twelve autistic children, using synchronous (movement-contingent) and asynchronous (non-contingent) sound mappings as a testbed to elicit contrasting interaction dynamics. Our results show that SKEC was sensitive to condition differences, whereas PC was not in this setting, indicating that OSMoSIS provides a suite of indices in which SKEC serves as the primary marker for dynamic, whole-body contexts and PC as a supporting, posture-aligned measure. Complementary coding with Interact behavioural coding software revealed reduced non-response under synchronous conditions, supporting the convergent validity of OSMoSIS. We argue that OSMoSIS advances the measurement of SMS beyond MEA, enabling analysis in ecologically valid contexts such as therapeutic and educational interventions.

As children face global challenges, creating environments that nurture hope and empower them to shape a fair, transparent future is essential.Conversational AI systems (CAIs) offer opportunities for cognitive and emotional growth, with trust built through transparent, responsive interactions.This workshop offers participants a hands-on opportunity to analyze child-CAI interactions, bringing their own use cases alongside pre-recorded examples from five countries provided by organizers.In collaboration with a diverse group of stakeholders, the focus will be on identifying the human factors that influence trust in child-AI interactions, aiming to advance guidelines for building transparent, trustworthy conversational AI systems.

In an era where Artificial Intelligence (AI) permeates our lives, its impact on children raises critical considerations. This workshop aims to delve into the multifaceted realm of Human-Centered AI (HCAI) for children, exploring the transformative role of AI in fostering creative expression and inclusive learning environments. Our goal is to unite diverse research expertise and methodologies, focussing on how AI can be tailored to meet diverse learning needs, enabling personalized and engaging educational experiences which support creativity. This workshop will bring together researchers, educators, technologists, and practitioners for expert talks, interactive demonstrations, and collaborative discussions,. Our goal is to foster a multidisciplinary dialogue on developing child-centered AI solutions that enhance creative learning while being mindful of inclusivity, ethical considerations and safeguarding against potential risks.

UPDATED—2024/04/17 13:30:54. This paper reports on work in progress to develop an assessment tool for Social Motor Synchrony (SMS) in children’s interactions. Expanding upon the foundation provided by the OSMoSIS motor sonification tool, we have developed an innovative approach for measuring SMS with children. Our method aims to overcome challenges stemming from variations in research settings and objectives encountered in studying SMS within the context of full-body interaction. Current tools for investigating SMS have been conducted in fixed settings, often limiting the interacting individuals to a seated position. However, this is in stark contrast to interventions such as music therapy, where the emphasis is on full-body interactions, and the settings exhibit considerable diversity. In such contexts, unrestricted movement is recognised as a fundamental element contributing to the overall efficacy and authenticity of the therapeutic experience, making it important to preserve the inherent benefits of a naturalistic setting. The study described in this paper explores the utility of OSMoSIS as a tool for assessing social motor synchrony revealing intriguing differences when comparing OSMoSIS with the well-known MEA software. This study not only brings a fresh perspective to the evaluation of social motor synchrony but also offers unique insights that set it apart from existing tools.

Assessing human vulnerability in cybersecurity is critical to understanding the relationship between human factors and the security of digital systems. This issue is exacerbated in areas such as public administration due to the sensitive nature of the data and services handled by government agencies, including citizen records, financial data, and national security details. Robust cybersecurity measures and understanding the human factors contributing to incidents are essential to securing public administration systems and maintaining public trust in government institutions. This poster presents ongoing work to develop a new psychometric tool to assess the human factors that play a critical role in cybersecurity incidents.

Interactional synchrony, the spontaneous coordination of movements during interaction, is increasingly considered important in research on the development of non-verbal communication by autistic children. There is evidence that interventions using embodied-interaction technologies to support interactional synchrony are possible, but we do not have a shared framework in Human-Computer Interaction (HCI) for designing and evaluating such systems. We discuss existing measurement and evaluation tools used in experimental psychology and consider how the prevalent approach could be adapted to naturalistic HCI study contexts, with input from domain experts. We report on an exploratory case study evaluating a full-body interactive musical system with a group of ten autistic children. We provide methodological recommendations for the evaluation of future systems focusing on interactional synchrony, highlight limitations of current measurement tools and suggest mitigations.

This paper describes the design, implementation, and pilot evaluation of an interface to support embodied musical interaction for children with Autism Spectrum Conditions (ASC), in the context of music therapy sessions. Previous research suggests music and movement therapies are powerful tools for supporting children with autism in their development of communication, expression, and motor skills. OSMoSIS (Observation of Social Motor Synchrony with an Interactive System) is an interactive musical system which tracks body movements and transforms them into sounds using the Microsoft Kinect motion capture system. It is designed so that, regardless of motor abilities, children can generate sounds by moving in the environment either freely or guided by a facilitator. OSMoSIS was inspired by the author's experiences as a music therapist and supports observation of Social Motor Synchrony to allow facilitators and researchers to record and investigate this aspect of the therapy sessions. It converts movements into sounds using Microsoft Kinect body tracking, in the context of an interactive game. From our preliminary testing with 11 children with autism (aged 5 – 11 years old), we observed that our design actively connects children, who displayed a notable increase in engagement and interaction when the system was used.

With a growing call for an increased emphasis on computing in school curricula, there is a need to make computing accessible to a diversity of learners. One potential approach is to extend the use of physical toolkits, which have been found to encourage collaboration, sustained engagement and effective learning in classrooms in general. However, little is known as to whether and how these benefits can be leveraged in special needs schools, where learners have a spectrum of distinct cognitive and social needs. Here, we investigate how introducing a physical toolkit can support learning about computing concepts for special education needs (SEN) students in their classroom. By tracing how the students' interactions-both with the physical toolkit and with each other-unfolded over time, we demonstrate how the design of both the form factor and the learning tasks embedded in a physical toolkit contribute to collaboration, comprehension and engagement when learning in mixed SEN classrooms.