This workshop addresses this gap by bringing together researchers and practitioners from AI, HCI, and the learning sciences to explore how interactive systems can better support learning. We focus on the design and evaluation of human-AI collaborative learning interfaces that are technically robust, human-centered, and pedagogically grounded. By fostering interdisciplinary dialogue, the workshop aims to identify shared challenges, design principles, and research directions for next-generation learning technologies.

Collaborative programming is a widely adopted classroom activity to encourage peer scaffolding, yet real-time collaboration often breaks down into parallel individual work with minimal interaction. Our formative studies reveal that even when students want to collaborate, they are held back by the effort required to understand a teammate's entire problem at once. We present Canary, a system that supports peer scaffolding by breaking down programming obstacles into smaller steps tailored to a student's skill level. Canary alerts potential helpers to specific places where they can start, using AI to turn complex problems into a step-by-step ladder that starts with easy fixes before moving toward harder logic. By providing this gradual ramp-up, Canary enables students to make quick contributions and progressively work toward solving their peers' problems. Our evaluation shows that this staged approach makes helping feel less overwhelming, leading to more frequent and effective collaboration among students.

Programming instructors have diverse philosophies about integrating generative AI into their classes. Some encourage students to use AI, while others restrict or forbid it. Regardless of their approach, all instructors benefit from understanding how their students actually use AI while writing code. Such insight helps instructors assess whether AI use aligns with their pedagogical goals, enables timely intervention when they find unproductive usage patterns, and establishes effective policies for AI use. However, our survey with programming instructors found that many instructors lack visibility into how students use AI in their code-writing processes. To address this challenge, we introduce Editrail, an interactive system that enables instructors to track students' AI usage, create personalized assessments, and provide timely interventions, all within the workflow of monitoring coding histories. We found that Editrail enables instructors to detect AI use that conflicts with pedagogical goals accurately and to determine when and which students require intervention.

As AI assistance becomes embedded in programming practice, researchers have increasingly examined how these systems help learners generate code and work more efficiently. However, these studies often position AI as a replacement for human collaboration and overlook the social and learning-oriented aspects that emerge in collaborative programming. Our work introduces human-human-AI (HHAI) triadic programming, where an AI agent serves as an additional collaborator rather than a substitute for a human partner. Through a within-subjects study with 20 participants, we show that triadic collaboration enhances collaborative learning and social presence compared to the dyadic human–AI (HAI) baseline. In the triadic HHAI conditions, participants relied significantly less on AI generated code in their work. This effect was strongest in the HHAI-shared condition, where participants had an increased sense of responsibility to understand AI suggestions before applying them. These findings demonstrate how triadic settings activate socially shared regulation of learning by making AI use visible and accountable to a human peer, suggesting that AI systems that augment rather than automate peer collaboration can better preserve the learning processes that collaborative programming relies on.

Code edit histories can offer instructors valuable insight into students’ problem-solving processes, revealing unproductive behaviors that final code alone cannot capture. For example, a correct solution may contain large copy-and-pasted segments (suggesting the code originated elsewhere) or unguided trial-and-error (suggesting a lack of clear strategy). Timelines are a common way to visualize code histories, but existing timeline visualizations of code or document histories show only when and where edits occurred, not what changed. Without this context, it is difficult to answer key questions about how students invested effort or to infer their intentions. We present CodeStream, a visualization system that augments timelines with situational code annotations, whose granularity and visibility dynamically adapt to scale and interaction state. A comparison study shows that CodeStream enables context-aware navigation of coding histories, supporting fast and accurate pattern identification, and helping instructors reason about students’ coding behaviors and identify who may need intervention.

University research labs often rely on chat-based platforms for communication and project management, where valuable knowledge surfaces but is easily lost in message streams. Documentation can preserve knowledge, but it requires ongoing maintenance and is challenging to navigate. Drawing on formative interviews that revealed organizational memory challenges in labs, we designed CHOIR, an LLM-based chatbot that supports organizational memory through four key functions: document-grounded Q&A, Q&A sharing for follow-up discussion, knowledge extraction from conversations, and AI-assisted document updates. We deployed CHOIR in four research labs for one month (n=21), where the lab members asked 107 questions and lab directors updated documents 38 times in the organizational memory. Our findings reveal a privacy-awareness tension: questions were asked privately, limiting directors' visibility into documentation gaps. Students often avoided contribution due to challenges in generalizing personal experiences into universal documentation. We contribute design implications for privacy-preserving awareness and supporting context-specific knowledge documentation.

Modern task-oriented chatbots present GUI elements alongside natural-language dialogue, yet the agent's role has largely been limited to interpreting natural-language input as GUI actions and following a linear workflow. In preference-driven, multi-step tasks such as booking a flight or reserving a restaurant, earlier choices constrain later options and may force users to restart from scratch. User preferences serve as the key criteria for these decisions, yet existing agents do not systematically leverage them. We present MAESTRO, which extends the agent's role from execution to decision support. MAESTRO maintains a shared preference memory that extracts hard and soft preferences from natural-language utterances and provides two mechanisms. Preference-Grounded GUI Adaptation applies in-place operators (augment, sort, filter, and highlight) to the existing GUI according to preference strength, supporting comparison among options. Preference-Guided Workflow Navigation detects conflicts between preferences and available options, proposes backtracking, and records failed paths to avoid revisiting dead ends. Through a controlled experiment (N=33), we demonstrated that MAESTRO improved decision quality in movie ticketing: final bookings left fewer hard preferences unmet, and users made fewer selections that violated their stated preferences during the process than in the baseline condition, although task success rate and completion time did not differ significantly. In addition, we showed that using MAESTRO in voice mode can increase users' active engagement as well as their mental burden, revealing the nuanced tension in agentic interaction design for conversational agents with a GUI.

AI programming tutors provide scalable support, yet lack the behavioral context human tutors rely on to adapt support to learners' needs. We present TutorTrace, a dataset and behavioral abstraction pipeline that makes learners' behavioral context visible and computable in real time from low-level IDE telemetry. Across four deployments in two introductory Python courses (N=480), TutorTrace captures approximately 180K telemetry events, 13,633 behavioral segments, and 27 continuously computed metrics. From this foundation, we derive a taxonomy of learner activity before the first AI query, between consecutive queries, and across the full session, enabling systems to respond not just to what learners say, but to what they have done leading up to the help-seeking moment. In a preliminary classroom evaluation, behavior-aware prompts were associated with a decrease in intervals between queries with no independent work from 50.0

When AI agents shift from answering questions to taking actions, users face a new problem: deciding what to delegate, to a system whose action space they cannot fully anticipate. We call the resulting dissatisfaction delegation regret, a pattern in which users regret not that the agent erred, but that it acted beyond what they would have authorized. In a controlled study, 20 university students completed five common daily tasks using OpenClaw, a general-purpose AI agent, across tasks chosen to vary in privacy, stakes, and reversibility. For each task we measured trust, perceived control, transparency, supervision burden, and approval preference on 5-point Likert scales, and collected free-text reflections analyzed through thematic coding. Three findings emerged. First, participants calibrated trust per task rather than per agent: they granted wide autonomy for advisory and low-stakes tasks but demanded confirmation for irreversible, externally visible actions. Second, irreversibility combined with external visibility, rather than stakes alone, appeared to drive trust withdrawal: the moderate-stakes email task triggered the sharpest drop in trust (M = 3.10) and the highest demand for approval (M = 4.65), whereas a high-stakes but verifiable task did not produce the same response. Third, delegation regret appeared consistently when the agent executed actions without preview, even when the output was rated as successful. We discuss implications for agent designs that expose action boundaries, support per-task autonomy policies, and separate advisory output from agentic execution.

Cognitive presence and social presence are crucial for a comprehensive learning experience. Despite the flexibility of asynchronous learning environments to accommodate individual schedules, the inherent constraints of asynchronous environments make augmenting cognitive and social presence particularly challenging. Students often face challenges such as a lack of timely feedback and support, an absence of non-verbal cues in communication, and a sense of isolation. To address this challenge, this paper introduces Generative Co-Learners, a system designed to leverage generative AI-powered agents, simulating co-learners supporting multimodal interactions, to improve cognitive and social presence in asynchronous learning environments. We conducted a study involving 12 student participants who used our system to engage with online programming tutorials to assess the system's effectiveness. The results show that by implementing features to support textual and visual communication and simulate an interactive learning environment with generative agents, our system enhances the cognitive and social presence in the asynchronous learning environment. These results suggest the potential to use generative AI to support student learning and transform asynchronous learning into a more inclusive, engaging, and efficacious educational approach.

Background and Motivation. Live coding is a common pedagogical technique where instructors write code in real time during lectures. For students, the main drawbacks of live coding are that it can feel too fast and it can be difficult to take notes. Objectives. Our work seeks to improve the student experience in live coiling lectures by: (1) understanding how instructors expect students to take notes and what challenges students face in doing so; and (2) investigating whether a specialized note-taking tool can help students keep up with the pace of the lecture and take better notes. Methods. Based on interviews h instructors who use live coding (n=10), we designed a simple note -taking interface consisting of a rich text editor Which allows students to take snapshots of the instructor's code. We conducted a within-subjects lab experiment (n=57) comparing our interface with a traditional code editor during two 15 -minute live coding lectures. We used quizzes and surveys to assess learning, mental workload, and student perceptions, and analyzed students' notes to determine how much information was captured from the lecture. Findings. In the experimental condition, NASA-TLX surveys indicated a significantly lower mental workload and students reported that they could more easily keep up with the lecture. Additionally, students perceived their notes to be more useful and our analysis revealed that the notes had significantly more information from the lecture and provided more context for copied code. Despite these benefits, we did not see a significant difference in learning between the two conditions. Implications. Our results show that during live coding lectures, we can decrease student mental workload arid increase the quality of notes by providing an interface which (1) allows capturing the instructor's code without having to type it out; and (2) maintains a clear visual distinction between code snippets and other text. Future work may examine if such an interface can lead to learning gains over long-term use in the classroom.

Modern organizations frequently rely on chat-based platforms (e.g., Slack, Microsoft Teams, and Discord) for day-to-day communication and decision-making. As conversations evolve, organizational knowledge can get buried, prompting repeated searches and discussions. While maintaining shared documents, such as Wiki articles for the organization, offers a partial solution, it requires manual and timely efforts to keep it up to date, and it may not effectively preserve the social and contextual aspect of prior discussions. Moreover, reaching a consensus on document updates with relevant stakeholders can be time-consuming and complex. To address these challenges, we introduce CHOIR (Chat-based Helper for Organizational Intelligence Repository), a chatbot that integrates seamlessly with chat platforms. CHOIR automatically identifies and proposes edits to related documents, initiates discussions with relevant team members, and preserves contextual revision histories. By embedding knowledge management directly into chat environments and leveraging LLMs, CHOIR simplifies manual updates and supports consensus-driven editing based on maintained context with revision histories. We plan to design, deploy, and evaluate CHOIR in the context of maintaining an organizational memory for a research lab. We describe the chatbot's motivation, design, and early implementation to show how CHOIR streamlines collaborative document management.

Facilitating class-wide debriefings after small-group discussions is a common strategy in ethics education. Instructor interviews revealed that effective debriefings should highlight frequently discussed themes and surface underrepresented viewpoints, making accurate representations of insight occurrence essential. Yet authoring presentations in real time is cognitively overwhelming due to the volume of data and tight time constraints. We present Dynamite, an AI-assisted system that enables semantic updates to instructor-authored slides during live classroom discussions. These updates are powered by semantic data binding, which links slide content to evolving discussion data, and semantic suggestions, which offer revision options aligned with pedagogical goals. In a within-subject in-lab study with 12 participants, Dynamite outperformed a text-based AI baseline in content accuracy and quality. Participants used voice and sketch input to quickly organize semantic blocks, then applied suggestions to accelerate refinement as data stabilized.

AI programming tools enable powerful code generation, and recent prototypes attempt to reduce user effort with proactive AI agents, but their impact on programming workflows remains unexplored. We introduce and evaluate Codellaborator, a design probe LLM agent that initiates programming assistance based on editor activities and task context. We explored three interface variants to assess trade-offs between increasingly salient AI support: prompt-only, proactive agent, and proactive agent with presence and context (Codellaborator). In a within-subject study (N=18), we find that proactive agents increase efficiency compared to prompt-only paradigm, but also incur workflow disruptions. However, presence indicators and interaction context support alleviated disruptions and improved users' awareness of AI processes. We underscore trade-offs of Codellaborator on user control, ownership, and code understanding, emphasizing the need to adapt proactivity to programming processes. Our research contributes to the design exploration and evaluation of proactive AI systems, presenting design implications on AI-integrated programming workflow.

Programming instructors often conduct collaborative learning activities, such as Peer Instruction (PI), to enhance student motivation, engagement, and learning gains. However, the impact of group discussion and formation mechanisms on student performance remains unclear. To investigate this, we conducted an 11-session experiment in a large, in-person CS1 course. We employed both random and expertise-balanced grouping methods to examine the efficacy of different group mechanisms and the impact of expert students' presence on collaborative learning. Our observations revealed complex dynamics within the collaborative learning environment. Among 255 groups, 146 actively engaged in discussions, with 96 of these groups demonstrating improvement for poor-performing students. Interestingly, our analysis revealed that different grouping methods (expertise-balanced or random) did not significantly influence discussion engagement or poor-performing students' improvement. In our deeper qualitative analysis, we found that struggling students often derived benefits from interactions with expert peers, but this positive effect was not consistent across all groups. We identified challenges that expert students face in peer instruction interactions, highlighting the complexity of leveraging expertise within group discussions.

This paper introduces REVA, a human-AI system that expedites instructor review of voluminous AI-generated programming feedback by sequencing submissions to minimize cognitive context shifts and propagating instructor-driven revisions across semantically similar instances. REVA introduces a novel approach to human-AI collaboration in educational feedback by adaptively learning from instructors' attention in the review and revision process to continuously improve the feedback validation process. REVA's usefulness and effectiveness in improving feedback quality and the overall feedback review process were evaluated through a within-subjects lab study with 12 participants.

This paper introduces SPHERE, a system that enables instructors to effectively create and review personalized feedback for in-class coding activities. Comprehensive personalized feedback is crucial for programming learning. However, providing such feedback in large programming classrooms poses significant challenges for instructors. While Large Language Models (LLMs) offer potential assistance, how to efficiently ensure the quality of LLM-generated feedback remains an open question. SPHERE guides instructors' attention to critical students' issues, empowers them with guided control over LLM-generated feedback, and provides visual scaffolding to facilitate verification of feedback quality. Our between-subject study with 20 participants demonstrates SPHERE's effectiveness in creating more high-quality feedback while not increasing the time spent on the overall review process compared to a baseline system. This work contributes a synergistic approach to scaling personalized feedback in programming education, addressing the challenges of real-time response, issue prioritization, and large-scale personalization.

Developing embedded systems is a complex endeavor that frequently requires collaborative teamwork. With the rise of freelance work and the global shift towards remote work, the need for effective remote collaboration has become crucial for many developers and their clients. However, current communication and coordination tools are predominantly tailored for software development rather than hardware-focused tasks. This study investigates the potential for remote support tools specifically designed for embedded systems development. Through interviews with 12 experienced embedded systems developers, we explored their existing remote work practices, challenges, and requirements. We also conducted a user enactment study featuring a custom-designed remote manipulation agent, Handy, as a theoretical assistant, to identify the kinds of support developers would value in a collaborative setting. Our findings highlight the scenarios and strategies employed in remote work, the specific support needs, and the challenges related to information exchange, coordination, and execution. Additionally, we explore concerns around privacy, control, and trust when using remote physical manipulation tools. This research contributes to the field by integrating the development of embedded systems with the remote, on-demand collaboration and assistance typical of software environments, offering a solid empirical foundation for future research on remote manipulation agents in this area.

Supporting novice computer science students in learning the software development life cycle (SDLC) at scale is vital for ensuring the quality of future software systems. However, this presents unique challenges, including the need for effective interactive collaboration and access to diverse skill sets of members in the software development team. To address these problems, we present ''DevCoach'', an online system designed to support students learning the SDLC at scale by interacting with generative agents powered by large language models simulating members with different roles in a software development team. Our preliminary user study results reveal that DevCoach improves the experiences and outcomes for students, with regard to learning concepts in SDLC's ''Plan and Design'' and ''Develop'' phases. We aim to use our findings to enhance DevCoach to support the entire SDLC workflow by incorporating additional simulated roles and enabling students to choose their project topics. Future studies will be conducted in an online Software Engineering class at our institution, aiming to explore and inspire the development of intelligent systems that provide comprehensive SDLC learning experiences to students at scale.