Building Instructional AI Assistants with TritonGPT: An Exploratory Pilot: This session presents the UC San Diego IT Services exploratory pilot of offering instructional AI assistants to faculty and students using TritonGPT, our secure campus-hosted generative AI platform. Drawing from a successful pilot at San Diego State University, we’ll showcase a two-phase approach: a faculty-facing assistant that streamlines content creation, quiz generation, and lecture material summarization, and a student-facing “virtual TA” that guides learners through course concepts using instructor-approved materials. The project emphasizes instructor control, transparency, and data privacy, with curated integrations into tools like Google Drive and Kaltura Media Gallery. Attendees will gain insight into implementation logistics, ethical considerations, and how locally hosted GenAI tools at the San Diego Supercomputer Center can assure data privacy while experimenting with AI-enhanced teaching. This presentation is ideal for those curious about responsible, practical GenAI deployment in academic settings.

Effects of a Chemistry Learning Intervention on Biology Student Learning and Experience: Students enter introductory biology with varying prior chemistry coursework. As chemistry concepts are foundational for introductory biology, differences in prior coursework could impact introductory biology course performance and students’ experiences learning chemistry in biology concepts. In this study we surveyed students and identified significant relationships between prior chemistry coursework, student background, and the likelihood of students struggling with chemistry and negative experiences learning biology. Additionally, we identified significant equity gaps on exams related to students' prior chemistry coursework and backgrounds. However, instructors are limited in the support they can provide during in-person class time during a 10-week quarter. As such we designed, implemented, and assessed a targeted chemistry intervention with a goal of supporting student learning of six chemistry in biology topics. The asynchronous intervention was delivered online via our course learning management system and consists of a mix of supplemental short videos and practice problems, as well as pre- and post- assessments of student understanding. We found that the intervention increased student understanding of chemistry and exam performance, but did not affect students’ experiences learning the material (Mendez et al., 2024, Meaders et al., 2024). In this presentation we will share our broad research findings, the structure of the intervention, and our tips for implementation.

LLMs feedback on student writing at scale: applications and lessons learned: Feedback is essential for learning, but providing it often requires difficult tradeoffs between speed, quality, and personalization. These considerations often inform not only approaches to feedback provision, but also how we choose what to assign. Large class sizes and short academic quarters intensify these challenges. LLMs offer solutions to some of these challenges but raise a host of new ones. This talk is about my experience developing an LLM-based tool to provide feedback on structured writing assignments. I will describe my approach, offer examples of its output, and share a series of lessons that I learned in developing it. Lessons include: the importance of model selection, the role and limitations of browser-based interfaces vs. APIs, how to write, evaluate, and revise prompts, and consideration of ethics and student perceptions. My implementation relies on R code and the Anthopic API. Those who wish to adapt this approach for their own use will need familiarity with (nearly any) programming language.

Custom Chatbots Made Simple: A No-Code Framework for Active Learning: In this session, we will demonstrate an intuitive, no-code chatbot template that empowers educators to rapidly create customized, AI-driven learning tools specific to their courses. This template enables faculty—without any programming experience—to quickly develop diverse chatbot-based activities suitable for any academic discipline. Attendees will experience live demonstrations of how custom chatbots from this template can effectively support asynchronous student learning, showcasing activities from both lower-division Biology courses and asynchronous interdisciplinary pedagogy workshops. These activities utilize the chatbot to engage students in digging deeper into content, learning computer code, studying, and interacting with quantitative simulations. Drawing from the implementation of the chatbots in multiple lower-division Biology courses, we will share student survey data that illustrates significant positive outcomes in students’ self-efficacy, engagement, and technological competencies, better preparing them for a rapidly evolving professional landscape. This session aligns with the themes of Artificial Intelligence and Online Learning by highlighting a practical, accessible strategy for leveraging AI to drive educational innovation. Participants will gain actionable ideas for using custom chatbots in activities to support student engagement and clear next steps to develop their own tailored chatbots.

Evaluating Pre-Service Teachers with Multimodal AI: Toward Objective and Scalable Presentation Assessment : Assessing the presentation skills of pre-service teachers is critical in teacher education, but traditional methods often suffer from subjectivity, inconsistency, and limited scalability. To address these challenges, we present a multimodal AI-based framework that evaluates presentation performance using audio, text, and visual data from micro-teaching sessions of 90 undergraduate students. The system extracts acoustic and linguistic features from speech, and facial expressions, gaze, and head pose from video. Correlation analysis revealed that features such as faster speaking speed and greater lexical diversity positively correlated with instructor scores, while longer sentences and sadness expressions were negatively associated. Multimodal classification using AdaBoost achieved 82.2% accuracy, outperforming unimodal approaches. This research aligns with the Expo’s themes of “Teaching with Tech” and “Artificial Intelligence” by demonstrating how AI can provide consistent, scalable feedback in teacher preparation. By identifying strong and weak presentations objectively, the system supports instructors in improving instructional feedback. Participants will learn how multimodal data can enhance assessment, review technical components such as feature extraction and classification, and explore how AI can support formative evaluation in teacher education.

Community-Centered Curriculum in Action: How the Inaugural SDUSD APIDA Student Board is Cultivating County-Wide Changemakers: In its inaugural year, this presentation highlights the launch of a culturally responsive, community-driven online program led by the inaugural APIDA (Asian Pacific Islander Desi American) Student Board, an initiative between College Corps at UC San Diego and the San Diego Unified School District Ethnic Studies and Youth Advocacy departments. Designed for APIDA middle and high school students and allies, the program combines Ethnic Studies frameworks, mentorship, and identity-affirming workshops delivered in a virtual format. Research shows that culturally relevant pedagogy increases engagement and academic confidence among historically underrepresented students (Ladson-Billings, 1995; Hammond, 2015). This board uses online platforms not just for content delivery, but to foster community, leadership, and self-reflection for youth in the second-largest school district in California. Aligned with the Online Learning and Teaching with Tech themes of the Education Innovation Expo, this session explores how technology can advance educational equity and student empowerment. Attendees will learn about culturally competent curriculum design, student engagement strategies, and the challenges and successes of launching this coalition in its first year. Key takeaways include practical tools for inclusive virtual teaching, insights on building school-university partnerships, and models for centering cultural identity in digital learning spaces through centering personal narratives.

A student-led, integrative case-based approach for teaching value-based medicine in preclinical medical education: Rising healthcare costs underscore the need to train future physicians in value-based medicine (VBM), which balances quality care with cost-conscious decision-making. In our pedagogical innovation, we evaluated a student-led, technology-enabled VBM intervention for second-year medical students at UCSD School of Medicine, focusing on deep vein thrombosis (DVT) diagnosis using the interactive HealthValueLab (HVL) case study platform. 92 students completed pre- and post-intervention surveys assessing knowledge, attitudes, and confidence for VBM and DVT diagnostics, with test-ordering behaviors analyzed before and after a VBM-focused lecture. Post-intervention, students reported significantly increased confidence and awareness of cost-conscious care (p < 0.05); however, changes in clinical decision-making, as measured by test selection in the HVL case, were minimal. Knowledge scores improved from 58.7% to 63.4%, though this was not statistically significant (p = 0.057). These findings suggest that a single-session, student-led digital intervention can enhance self-perceived preparedness for VBC, but is insufficient to meaningfully alter clinical decisions. The HVL platform provided valuable insights into student attitudes and behaviors, highlighting the need for novel, iterative educational strategies. Our results indicate that while digital, student-centered tools effectively engage learners, ongoing reinforcement is essential to translate VBM principles into clinical practice.

Personalized Hints Based on Student Sketches: Sketching has been part of education for over a hundred years, but has been largely overlooked by Ed Tech which mostly relies on multiple-choice and numerical entry. We have developed software that automatically grades students freehand sketch for spatial visualization training. The software provides personalized feedback to encourage persistence and productive struggle. This approach has been shown to increase student performance in spatial visualization as well as ability to communicate with teammates in an introduction to engineering design course. Currently we are evaluating student learning from sketch to sketch to see how personalized feedback can be improved. We are also using student sketches to generate personalized email to send to students to encourage them when needed and acknowledge increased performance when it occurs. Come see our latest results!

Massively-scalable auto-graded online assessments with adaptive feedback using PrairieLearn: Resulting from multiple factors (larger enrollments, increased online course offerings, technological advances, etc.), auto-graded online assessments have become increasingly critical in supporting modern university courses. Recent advances in online assessment platforms have enabled the design and incorporation of immediate adaptive feedback that is uniquely tailored to a student's specific misconceptions. In this presentation, I will discuss how to design and develop such online assessments using the PrairieLearn platform, and I will demonstrate how I have utilized PrairieLearn to implement massively-scalable assessments in 3 vastly different courses: CSE 100 (Advanced Data Structures in C++; hybrid course), CSE 6R (Intro Python for Non-CSE Majors; online course), and CSE 185 (Advanced Bioinformatics Lab; in-person course). I will also demonstrate how, when utilized with PrairieTest via the Triton Testing Center, I have been able to reuse the assessments I developed directly within self-scheduled proctored in-person exams.

Making Computer Science Concepts Visible and Virtually Tangible through Creative Coding in Virtual Reality: Embodied Code (EC) is a creative coding toolkit in virtual reality (VR). It offers novice coders a low barrier to entry for exploring fundamental computing concepts in playful ways using a node-based visual programming language (VPL). VPLs have served as an alternative to text-based programming for decades – however, EC transforms visual programming into an immersive experience, allowing users to enter the world where they are actively coding. To create a computer program, the user assembles the necessary nodes and wires them together to make larger execution trees. However, considerable flexibility is afforded in placing and manipulating the elements of code such that space and body movement can be leveraged as emergent organizational and conceptual scaffolds. This lightning lab demonstrates how spawning objects, changing their appearance, and adding game engine effects are accomplished within a project to construct a simple catapult. Opportunities for embodied learning through the arrangement of nodes and output are also discussed. Finally, preliminary work will be presented exploring how college students with little computer science experience can benefit from Embodied Code. It will be argued that fundamental computational concepts can be induced through analogical reasoning on the basis of exposure to a small number of examples.

Designing for the Future Now: Architectural Perspectives on AI and Educational Technology: This panel brings together UC San Diego and four practicing architects with deep experience in educational environments to explore how artificial intelligence and emerging technologies are reshaping the design of learning spaces. With campuses embracing hybrid pedagogy, digital collaboration, and immersive tools, this discussion will examine how architectural innovation can support—and be transformed by—these trends. This session will address four core questions: How is AI influencing the design of educational spaces? What spatial adaptations best support hybrid and asynchronous learning models? How can physical environments impact innovation in a tech-enhanced academic world? What design strategies anticipate the long-term integration of intelligent systems in learning environments? Panelists will share case studies from higher education projects, including UC San Diego and peer institutions, illustrating how flexible, responsive, and data-informed architecture can empower both pedagogy and campus life. The panel will also reflect on design ethics, and the evolving relationship between learners, faculty, and space. Attendees will gain insight into how interdisciplinary collaboration between educators and designers shapes more adaptable, technology rich AI-ready campuses. The discussion will offer both visionary and pragmatic takeaways for academics, technologists, and students interested in how physical spaces can catalyze educational transformation in an AI-driven future.

Active Learning in Organic Chemistry with 3D Printed Space-Filling Molecular Models: Ball-and-Stick molecular models are common aids used by students to help understand connectivity and Euclidean geometry when learning Organic Chemistry. Space-filling atomic models, on the other hand, allow the construction organic molecules with realistic dimensional and steric accuracy. However, model sets are rarely deployed during in-class exercises. As result many students do not know how to use the models and do not purchase the time-honored visualization aids. Here we address this issue and leverage 3D printing technology to introduce and teach students 3D structure, including sterics, and conformations of organic molecules through the in class use of 3D printed space-filling models. We will present 1) an accessible 3D printed version space-filling model set where low-cost, accurate atomic subunits can be printed out on any FDM 3D printer at large scale and 2) in class active learning exercises, using the new 3D Printed Space-Filling Molecular Model sets. Student engagement and feedback will be discussed alongside potential learning gains.

Reimagining DIY Video: Innovations in Studio U Design: At UC San Diego, Studio U has become a model for scalable, self-service content creation. In this short talk, we’ll showcase the newest innovations that are expanding how faculty and staff create professional, engaging media without needing on-site production crews. Recent enhancements include a two-person podcast/interview recording configuration that offers a polished, conversational setup ideal for instructional storytelling and guest engagement. A built-in teleprompter now supports more confident delivery of prepared scripts, especially for those new to on-camera work. Additionally, our latest integration allows users to bring in remote Zoom participants seamlessly, merging virtual presence with in-studio quality. These upgrades not only increase the versatility of the space but also lower the barrier to entry for high-quality educational video. Attendees will gain insight into the practical technology behind these features, lessons learned during implementation, and how these additions have helped drive user adoption and creative use cases across disciplines.

Classroom Simulacra: Building Contextual Student Generative Agents in Online Education for Learning Behavioral Simulation: Simulating virtual students in online learning offers new opportunities for educators to refine and innovate their teaching strategies. However, most existing approaches overlook how course materials dynamically modulate student online learning behaviors. In this presentation, we introduce a novel framework that addresses two key challenges: the lack of granular datasets for student simulation and the difficulty of modeling long, complex educational content. We share insights from a 6-week workshop involving 60 students, where we collected fine-grained behavioral data through a custom online education platform. Building on this, we present the Transferable Iterative Reflection (TIR) module, an AI technique that enhances large language models (LLMs) for more accurate student simulation. Our findings show that TIR outperforms traditional models, even with limited data, by better capturing learning dynamics over time. As such, this work of AI-powered student simulation in online learning is relevant to educational innovation at the intersection of Artificial Intelligence, Online Learning, and Teaching with Tech. Attendees will learn how AI-driven student simulation can create "digital twins" of classrooms, allowing for scalable, personalized education experiments and course optimization. We will also discuss how this approach opens new possibilities for data-driven instructional design, supported by our recent paper in ACM: https://dl.acm.org/doi/10.1145/3706598.3713773.

AI Literacy, It's Critical!: Generative AI (GenAI) tools present educational challenges and opportunities. They can replace or augment human thinking, take over or enhance problem-solving, hinder or amplify human creativity, and inhibit or accelerate learning. It all depends on how they’re used. Universities therefore have a responsibility to help students develop the ability to evaluate and engage with GenAI tools ethically and reflectively. Such AI Literacy is critical if we want students to continue to center themselves in the process of learning, doing and thinking, thereby equipping them to develop human, durable skills within an AI-infused world. To address this need, a group of UC San Diego professional staff from the Teaching & Learning Commons and Academic Integrity Office collaborated to develop a CANVAS module on Critical AI Literacy. This CANAVS module answers four central questions: 1) What is Artificial Intelligence, 2) How Does It Work, 3) How Do I Use it, and 4) How Can I Use it Responsibly? In the session, we will introduce the CANVAS module by facilitating an AI Literacy lesson with the participants interacting as students. By the end of the session, participants will have: 1) identified opportunities for incorporating AI Literacy in their department/courses; 2) critiqued their own AI Literacy to identify areas for growth or deepened learning; 3) analyzed the module for future improvements; and, 4) created a list of future projects to help our campus adapt to an AI enabled world.

Emergence-Driven Concept Maps: AI-Powered Knowledge Acquisition for Personalized Learning: We present the Knowledge Acquisition System (KAS), an AI-powered platform that uses an emergence-driven approach to personalize learning. Emergence, within this context, refers to the identification of non-obvious yet structurally significant connections within knowledge networks. KAS helps learners form deeper understanding through dynamic, inductive concept maps. A core emergence metric, grounded in network theory, highlights long-range, cross-disciplinary and nonlinear relationships that enhance learning. AI automates concept map generation from traditional materials (e.g., slides, textbooks), automatically identifying crucial connections and creating tailored pathways that guide learners through intricate concepts with enhanced clarity and efficiency. Pilot implementations with advanced mathematics students at UC San Diego confirmed KAS's potential to streamline curriculum preparation and enhance students' comprehension and engagement. Attendees will learn: • How emergent subgraphs enhance cognitive structure formation, uncover long-range conceptual links, and trigger “aha” moments • How AI can automate the construction of adaptive concept maps from existing course material • Practical steps to build AI-driven, personalized learning pathways

Cultivating Well-being and Self-Care: Empowering Learners Through Life Skills and Integrative Practices in Online Education: This presentation explores the integration of well-being and self-care education within online learning environments. It highlights the critical role of teaching life skills alongside practical self-care strategies. The relevance to educational innovation lies in addressing a significant gap in traditional online curricula, which often prioritizes academic content over holistic student development. This approach moves beyond purely cognitive learning to cultivate the socio-emotional competencies essential for thriving in both academic and personal lives. This topic aligns with the "Online Learning" theme by addressing the unique challenges and opportunities of fostering student well-being in virtual educational settings. It provides practical strategies and frameworks for educators and instructional designers to embed well-being and self-care into online course design and delivery. Participants will learn 1) how to incorporate life skills education into online curricula in engaging and accessible ways; 2) practical strategies and frameworks, including the FEMA (Find, Embrace, Move, Again) method and the Self-care Canvas, that are adapted for online learning environments, and 3) the importance of fostering a supportive online learning community that normalizes discussions around well-being and self-care.

Embodied AI: Medical English Roleplays with RIA: As AI transforms healthcare and education, language programs can harness emerging technologies for skills-based training. This presentation shares how DES' Medical English program integrated RIA, an AI-powered humanoid robot, marking the first use of robotic simulation in language instruction on campus. Designed for international healthcare professionals, the program emphasizes patient-provider communication, clinical terminology, and U.S. healthcare delivery. By programming RIA with customizable patient profiles — each with a unique backstory, personality, and symptoms — we created lifelike roleplay scenarios for students to practice medical interviews, diagnosis discussions, and treatment explanations in English. The session will outline the integration process, including curriculum adjustments, collaboration with robotics engineers, and student preparation. We’ll share sample scenarios, student feedback, and observed outcomes. Preliminary results suggest RIA helped students build fluency, manage unpredictable patient responses, and practice sensitive conversations in a low-risk setting. Attendees will gain practical ideas for integrating AI-driven tools into language instruction for healthcare. https://www.cbs8.com/article/tech/ai-powered-robots-acting-patients-train-doctors/509-1d110cff-48d0-4726-a4ae-f808dde03eb3

Building STEM Confidence at Home: Designing Online Hands-On Labs for Structural Engineering: How do you turn a traditionally tactile, project-based engineering course into a rigorous online experience, without losing the “hands-on” magic? In this session, we share our process for creating an online structural engineering course for high school students, complete with physical lab kits, team-based design projects, and weekly synchronous lab sessions. Developed through UC San Diego’s Discover initiative to increase access to college-level education for underrepresented students, this course emphasized equity, project-based learning, and student self-efficacy. We will walk through the development of our lab activities and kits, discuss the pedagogical frameworks that guided our choices, and share lessons learned on implementation, logistics, and student engagement. Topics include designing lab kits within budget constraints, integrating live lab sessions, and fostering community and persistence in a fully online environment. This presentation aligns with the Online Learning and Teaching with Tech expo themes. Attendees will leave with practical design strategies for creating engaging, hands-on online lab experiences, as well as implementation tips that can be adapted to a range of disciplines and student populations.

Human Anatomy in Virtual Reality: Virtual reality (VR) is designed to create repeatable, controlled, and immersive environments that are challenging to create in the real world. We'll demonstrate the potential of virtual reality in the classroom by exploring how VR will be used in an upcoming human anatomy course at UCSD. Participants will follow along as we interact with and fly through key bodily structures in full immersive 3D.

Prompting AI to replace my professors: This talk discusses the revolutionary role of Artificial Intelligence in optimizing and individualizing academic learning, illustrated through the example of my own experience as a UC San Diego business economics student. I have successfully replaced the convention of lecture class attendance by employing AI-based tools such as personalized exam prep guides, on-demand concept explanations, and individualized study calendars and have continued a great academic track record (average grade of A-). By carefully guided interactions with AI assistants, I improved effectiveness in the mastery of advanced econometrics (ECON120B) by employing step-by-step problem-solving as well as formula sheet-guided tutoring. The outcome was significant improvement in understanding, retention, and performance on exams without access to live lectures. It illustrates how AI can be a dynamic on-demand educational companion that adapts dynamically according to individual needs and enhances academic success. Participants will learn about effective ways of incorporating AI into daily studies, how AI supports independent learning competencies, and guidelines for ethical use. The presentation will directly correspond with the “Artificial Intelligence” theme of the Expo by pinpointing how AI can remodel pedagogy, improve productivity, and individualize the learning experience of contemporary learners.

Integrating and Evaluating LLM-integrated Pedagogies in Introductory Computing Courses: Background: Large Language Models (LLMs) have broad implications for education in general, impacting the foundations of what we teach and how we assess. This is especially true in computing, where LLMs tuned for coding have demonstrated shockingly good performance on assignments historically used in introductory CS (CS1) courses. Objectives: This study aims to design and evaluate a new CS1 course that integrates LLMs as learning tools, focusing on student outcomes and perceptions. Methods: A new CS1 course was developed at UCSD using design principles tailored for LLM integration. Student performance was evaluated through assessments, while perceptions of LLM tools were captured by surveys. Data analysis focused on both overall outcomes and demographic subgroup differences. Results and Conclusions: The findings indicate that student performance outcomes, including differences across demographic groups, were largely consistent with historical results from courses without LLM integration. Students generally found LLM tools beneficial, though some expressed concerns about over-reliance. This work directly relates to the artificial intelligence theme of the expo, as our team explores how AI-integrated pedagogical strategies influence student outcomes. Participants will learn about results from our AI-integrated course, and points of student struggle with LLM tool usage and how we have iterated in our course design as a result of these findings.

From Controller to a Classroom: How Gaming Adjacent Education has Revolutionlized UC San Diego: Gaming, Esports, and Game Adjacent Development tools have been heavily more prevalent in education sectors within the recent decade. Gaming, especially in research specific fields, have innovated and revolutionized multiple fields of pursuit. The adaptation of typical development tools has been able to make massive waves within CogSci, Healthcare, and Defense research spheres with their ability to create “Serious Games” to help spearhead new research efforts. Commercial Gaming and Esports have also been more incorporated in different fields of education for enhanced learning and enriched teaching opportunities. The Entertainment Software Association & the Network of Academic and Scholastic Esports Federation have published numerous articles that gaming has heavily increased engagement along with excelling learning capabilities, teamwork, and industry readiness with college level students. UC San Diego has started to become a major public university existing amongst this growing ecosystem. Multiple labs within the last few years have started to use “Serious Games” to excel CogSci and Simulation Research in Qualcomm. This has also been matched with a growing scene of commercial gaming on campus with new esports facilities, massive large scale industry involvements, and one of the largest collegiate gaming scenes in the country. Overall, the impact and implementation of these gaming adjacent aspects have vast capabilities for the future of innovating education at UC San Diego.

Reshaping Engineering Education for the GenAI Era: A Pre-/Post-Class Assessment Framework for Advancing Classroom Learning, Evaluating In-Class Assignment for Graduate Students Using Bloom’s Taxonomy-Based Pre and Post-Lecture Formative Assessments: As generative AI (GenAI) tools can now complete college-level homework, educators face a growing challenge: how to fairly and accurately assess student learning. This presentation shares a practical solution—AI-proof, in-class formative assessments—and examines their effectiveness in a graduate bioengineering course at UC San Diego. Grounded in Black and Wiliam’s formative assessment theory and Bloom’s taxonomy, we implemented short pre- and post-lecture self-assessments aligned with course learning outcomes. These low-stakes in-class activities provided timely feedback and revealed stronger correlations with student performance than traditional homework, despite not contributing to final grades. This suggests they more accurately reflect students’ independent understanding—free from GenAI tools. This work supports the Artificial Intelligence and Teaching with Tech themes by offering a proactive response to AI’s growing role in education. Instead of banning AI or increasing surveillance, we focus on redesigning assessments to uphold integrity and support authentic learning. Participants will learn: 1- Why in-class assessments are critical in the GenAI era. 2- How to align assessments with Bloom’s taxonomy and course goals. 3- How to use assessment data to adapt instruction and improve learning outcomes. By rethinking assessment strategies, we can ensure technology enhances rather than undermines education—supporting fairness, engagement, and continuous improvement.

Cultivating Engagement: Lessons from the library classroom: As a librarian who teaches single-session library research classes, I always look for activities to quickly build engagement with students. This was especially true during the initial pivot to online instruction during the COVID-19 pandemic. For librarians accustomed to teaching information literacy sessions in person, it was a scramble to adapt teaching methods to the online environment while preserving an active learning approach to encourage learner-teacher interaction. As an instruction librarian, I experimented with various techniques to engage with students after the pivot to online instruction. As more library instruction sessions return to face-to-face or hybrid modes, what can we take from these creative methods to increase participation and engagement in the physical classroom? How can we pivot and use the multimodal activities we experimented with to increase student participation in face-to-face instruction? This presentation will introduce some techniques adapted from online instruction to engage with students in the in-person classroom, touching on both online instruction and teaching with tech. Participants will learn simple ways to incorporate low-stakes active learning methods into online and in-person classrooms to make instruction more engaging.

Fostering Interdisciplinary AI Education Through Project-Based Learning: This presentation shares a case study on teaching artificial intelligence (AI) through a project-based learning (PBL) framework designed to support diverse, interdisciplinary learners. Conducted at UCSD and through Project Nous, a nonprofit promoting liberal arts–style education in underserved communities, the seminar introduced AI to high school graduates by combining theory with hands-on design projects. Students created personalized AI systems—from emotion recognition to autonomous driving—while exploring core topics like machine learning, neural networks, and AI ethics. The curriculum follows five iterative phases: problem definition, data collection, model building, prototyping, and interpretation. Learners used tools like TensorFlow and PyTorch while reflecting on AI’s real-world impact. This model is adaptable to other AI-related courses, including hybrid or online formats, and serves as a testbed for educational technologies. Attendees will learn: • How PBL supports interdisciplinary AI education • How to scaffold without requiring prior expertise • Ideas for modular, tech-enhanced curricula and creative project design

JOPRO & the New Learning Ecosystem: Designing Tools and Communities to Foster Contextual, Developmental, and Trajectory-Aware Learning in Inter-/Un-disciplinary Spaces: JOPRO is an interdisciplinary incubator reimagining how early-stage researchers, educators, and changemakers engage with learning, leadership, and AI. This presentation explores JOPRO as a scalable model for fostering strategic mentorship and context-aware education—equipping individuals to navigate the complexities of innovation across fields and institutions. We will introduce FrontierMap, a new AI-augmented learning tool under development that helps users build historical, conceptual, and epistemic “maps” of emerging fields. By combining generative AI with structured frameworks for key figures, events, themes, controversies, and schools of thought, FrontierMap empowers learners to orient themselves at the edge of knowledge. We’ll also feature PlotTwisters, a narrative-based edtech initiative teaching co-regulation, self-awareness and critical inquiry of narratives, and Orthogonal Research & Education Lab, an open-source & open-science collective fostering transdisciplinary research and digital pedagogy. This session will engage participants in the challenges and opportunities of designing learning experiences that are interdisciplinary, AI-augmented, and context-sensitive. Attendees will leave with a deeper understanding of how to support meaningful learning and innovation in an age of complexity—and how to build tools and communities that respond dynamically to it.

"GenAI Tools for the Classroom: Streamlining Oral Exams and Student Data Analysis: This lightning lab showcases two practical open-source tools to address real classroom challenges using Generative AI. First, the "Scaling Oral Exams" application addresses the time-consuming nature of oral assessments. Instead of scheduling individual meetings or relying solely on written tests, instructors can pose questions that students answer verbally using their webcam. The system automatically transcribes responses and provides GenAI-assisted grading with a regrade request feature. The second tool, "Radiant Data Analysis Assistant," helps students analyze datasets without requiring advanced programming skills. Users simply upload their data (CSV/Parquet files) and type natural language requests like "estimate a model with price as the response variable." or "Is demand larger than 1750?". The assistant selects the appropriate tool(s), generates python code, runs the analysis on a server or local system, and explains the results in clear, understandable terms. This makes data literacy accessible to students across disciplines, not just those with strong technical backgrounds. Attendees will see these tools in action and explore practical considerations for adopting similar GenAI-assisted approaches in their own teaching.

Thinking Beyond Slides: Innovative Course Media Approaches: : Multimedia Services and Digital Learning will share examples from recent online/hybrid/flipped courses and discuss video approaches that best take advantage of the medium from a technical and pedagogical perspective.

GenAI in Programming Pedagogy and Assessment: The integration of generative AI (GenAI) into programming instruction and assessment offers numerous educational and practical benefits for both students and instructors. These include instant feedback, offloading of lower-order tasks, and improved management and workflows. However, this integration also comes with technical challenges and academic integrity and equity issues that need to be carefully addressed to ensure fair and effective learning and testing environments. This presentation describes the integration of GenAI into an introductory programming course for engineering students, outlining the instructional adaptations that supported its adoption. These included the redesign of lectures, assignments, and assessments for alignment with AI-era learning. The revised pedagogy emphasized critical thinking, higher-order programming tasks, metacognitive awareness, self-regulated learning, collaborative programming, and AI literacy. This encompassed high-stakes assessments, which were restructured in alignment with the principles of authentic assessment and allowed the use of AI tools by students, reflecting real-world working conditions. The presentation will offer insights into the technical challenges and educational trade-offs involved in integrating GenAI into a programming course, including its assessments.

Democracy is Digital: Cultivating Engaged Citizens Through Digital Media: Today, democracy lives and breathes in digital spaces. From local activism to global movements, digital media has become the primary way information is shared, civic identities are formed, and collective action is organized. In a world where knowledge and influence are increasingly mediated through digital channels, equipping students with the skills to critically engage and communicate through these platforms is essential for all disciplines and professions. As educators, we have a responsibility to prepare students for the realities of a digital age, ensuring they can navigate, critique, and contribute to the conversations that shape their lives. This presentation explores how digital media campaigns can be used as a powerful tools for teaching students about society issues, democratic practices, and public discourse. Pulling from examples from my own courses, I will demonstrate how utilizing digital media projects fosters civic agency, sharpens critical thinking and communication, and prepares students for professional and civic life in a rapidly changing media landscape.

In the age of AI, (human) tutors are more essential than ever: Traditionally, students have relied on tutors (or TAs) for help with homework and understanding concepts. However, in the age of generative AI, students often seek help from always-available GenAI systems instead, leaving tutor and office hours empty. Rather than concluding that human support for learning is no longer needed, I saw this as an opportunity to make better use of the personalized human support tutors and TAs can provide. In this presentation, I will discuss a framework for leveraging human tutors to provide personalized, proactive support for students, diverting these valuable human resources to the students who most need it and are least inclined to seek it out.

Feeding Two Birds with One Scone: Using AI to Help Students and Teachers: Traditional student papers are often limited by difficulties in formulating questions, identifying and mastering relevant literature, developing original ideas, writing clearly and grammatically, applying rubrics, and incorporating instructor comments. To address these challenges, we introduce an innovative educational model that reimagines and expands the Three-Dimensional Learning Assessment Protocol (Laverty 2016). In our model, instructors benefit from AI-assisted formative assessments that enhance learning, manage AI-generated content, and reduce feedback bias. We propose integrating large language models, such as ChatGPT, into (a) student research workflows, e.g., question formulation, methodology, literature analysis, and writing—and (b) instructional practices, including the development of lesson plans and assessments grounded in standard pedagogies, as well to manage assessment data and streamline their pedagogical practices. The approach is designed to be low-barrier: it uses freely available tools, fits within existing course structures, and requires low technical expertise to implement. Attendees will gain a practical, replicable model for using AI to enhance student writing and support instructor assessment in writing-intensive courses. Our presentation aligns with the themes of AI and Teaching with Technology. Our proposal lays the foundation for developing tailored AI-based tools that align with instructors' goals and students' learning objectives.

Practical Prompting Practice in AI Tools: This interactive session explores the practice of prompt engineering, a crucial skill for harnessing the power of AI in learning environments. Learn how to craft effective prompts that elicit desired responses from AI models, enhancing teaching and learning outcomes. Understanding these concepts will help us in determining how AI can support our goals, whether we are educators or learners.

Assessment for Learning: Practical Approaches using Quiz-Based Tools: The integration of digital tools into Kazakhstan’s higher education system presents both opportunities and challenges as the country modernizes its academic landscape. While platforms like Moodle and Google Classroom have become commonplace, especially after the pandemic, significant disparities remain in technological access between urban and rural institutions. However, challenges persist. Limited infrastructure in rural areas hinders equal access, while some educators resist adopting new technologies due to lack of training or preference for traditional methods. Additionally, students may experience digital fatigue from excessive screen time. To overcome these barriers, Kazakhstan must invest in broadband expansion, provide ongoing teacher training, and develop blended learning models that combine digital and in-person instruction. Government initiatives like Digital Kazakhstan should prioritize equitable technology distribution, while universities could partner with edtech companies to pilot innovative solutions. By addressing these challenges strategically, Kazakhstan can harness digital tools to create a more inclusive, effective higher education system. The future of learning lies in balancing technological advancement with human-centered teaching approaches, ensuring all students benefit from the digital revolution in education.

Computer-Based Testing at UC San Diego’s Triton Testing Center: The Triton Testing Center’s presentation will be on UC San Diego’s innovative Computer-Based Testing Facility (CBTF) which uses the software PrairieLearn. This secure facility promotes student learning with advanced testing software that can assess students’ abilities in a variety of disciplines such as mathematics, computer science, engineering, physics, analytical writing, business management, and political science. The PrairieLearn software can create robust, randomized tests that promote mastery of the material and integrity among students by ensuring that each student receives a different version of the test. Chen et al. (2018) found that “increasing the size of the pool from which problem generators are selected lessened the score advantage of collaborative cheating,” and that “randomization is an effective tool to discourage collaborative cheating… for asynchronous computerized exams.” (pg. 8). The combination of PrairieLearn’s capabilities and the Triton Testing Center’s secure environment creates the optimal conditions for fair, unbiased assessments. Students benefit from mastery-based testing, real-time feedback, and frequent, lower stakes testing. Leveraging the CBTF, instructional teams gain valuable time that they can devote to instruction, support for students, and require less resources to host and grade exams. Participants in this presentation will learn about the CBTF and receive a demonstration of how the PrairieLearn software can support student learning.