Key Takeaways
- The One‑U Responsible AI Initiative hosted its second annual “Minds and Machines: AI in Education” hackathon over 48 hours in mid‑September, drawing 102 participants from 26 teams.
- The event attracted entrants from industry and six other universities, highlighting its broad appeal beyond the University of Utah campus.
- Seven challenge tracks, enablement sessions, and expert meet‑ups supported teams as they built AI solutions for real‑world classroom problems.
- First‑place team Mindstone’s AccessLens delivers real‑time, multimodal accessibility aids; second‑place Patient Zero offers an AI‑driven standardized patient for health‑profession training; third‑place Team Dori creates collaborative, IEP‑aligned activities for elementary learners.
- Honorable‑mention projects ranged from adaptive assignment platforms and AAC communication bridges to quality‑assurance tools for AI tutoring systems, each receiving cash and AWS‑credit prizes.
- Judges emphasized equity, practical impact, and the importance of keeping AI as a supportive tool rather than a decision‑making authority in educational settings.
Overview of the Hackathon
The One‑U Responsible Artificial Intelligence Initiative’s “Minds and Machines: AI in Education” hackathon took place over a packed 48‑hour window in mid‑September. As the initiative’s second annual event, it invited anyone interested—though it was explicitly geared toward students—to apply data science, machine learning, and AI techniques to genuine classroom challenges. According to the organizers, the hackathon “brought together 102 participants across 26 teams to leverage data science, machine learning and AI to tackle real‑world issues in the classroom.” This concise statement captures the core mission: translating cutting‑edge technology into practical educational tools.
Participation and Geographic Reach
While rooted at the University of Utah, the hackathon’s draw extended far beyond campus borders. The registration tally topped 165 initial sign‑ups, ultimately yielding 102 active participants representing seven universities, industry professionals, and independent enthusiasts. The event’s brochure highlighted that it “reached well beyond the U, attracting participants from industry and six other universities.” Such diversity enriched the problem‑solving environment, allowing perspectives from K‑12 educators, medical trainers, and software engineers to converge on shared goals.
Structure: Challenge Tracks, Sessions, and Expert Guidance
To steer teams toward viable solutions, the hackathon featured three distinct challenge tracks, four enablement sessions, and a subject‑matter‑expert meetup. These components provided both thematic focus and hands‑on support, ranging from workshops on data privacy in AI to mentoring on model deployment. The schedule also allocated time for teams to consult with 13 expert judges, whose backgrounds spanned education, clinical simulation, and AI ethics. This structured scaffold ensured that participants could move from concept to prototype within the tight 48‑hour window.
First Place: AccessLens by Team Mindstone
The top honor went to Team Mindstone for their project AccessLens, described as “a real‑time accessibility assistant that transforms one instructor‑approved lesson feed into synchronized visuals, structured text, captions, audio, dyslexia‑friendly formatting, and spatial AR.” In their own words, AccessLens “bridges the gap between formal accommodations and the moment instruction happens while giving students private control over how they learn.” By delivering multimodal content on demand, the tool addresses the lag often experienced between a teacher’s lesson plan and the individualized supports that students with disabilities require. The judges praised its potential to promote inclusivity without singling out learners in a classroom setting.
Second Place: Patient Zero
Claiming second place, the team behind Patient Zero presented an AI‑powered standardized patient for health‑profession students. As the project summary notes, “A student interviews a talking virtual patient, then receives a scored report graded against real clinical guidelines. A built‑in equity check re‑runs the identical case across patient demographics and shows whether the student’s accuracy holds.” This dual focus on clinical skill assessment and equity monitoring aims to ensure that future clinicians not only master diagnostic techniques but also recognize biases that could affect patient care. The virtual patient’s ability to adapt scenarios based on race, gender, age, and socioeconomic status offers a scalable way to embed fairness into medical training curricula.
Third Place: Dori
Team Dori secured third place with a platform aimed at elementary students who have Individualized Education Programs (IEPs) or 504 plans. According to their description, “Dori provides elementary students with disabilities who have IEPs and 504 plans with accessible and collaborative learning activities directly under the supervision of their parents/guardians and teachers.” The system emphasizes “clear collaboration between teachers, students and families” and treats every activity as shared evidence that keeps the IEP current. Importantly, Dori refrains from making high‑stakes decisions autonomously, positioning AI as a facilitative aid rather than a replacement for human judgment—a design choice that resonated with the judges’ emphasis on responsible AI use.
Honorable Mention: Mithril: Assign
The honorable‑mention project Mithril: Assign introduces an adaptive assignment platform that moves beyond single‑score grading. As the team explains, “The teacher defines what a student must learn and demonstrate; Assign first tests the students, then iteratively finds the gap behind each wrong answer, reteaches it from a new angle, then verifies it with fresh questions. Teachers get concept‑level evidence of each student’s mastery instead of a single score.” By pinpointing specific misconceptions and offering targeted reteaching loops, Assign seeks to transform assessment into a diagnostic, formative process that informs both instruction and student self‑regulation.
Honorable Mention: Pebble
Another notable entry, Pebble, tackles communication barriers faced by nonverbal students. The project summary states, “Pebble is a real‑time AI bridge that transforms button taps on an augmentative and alternative communication speech board into natural spoken sentences. Over 10 million nonverbal individuals globally, including 1 in 40 K‑12 students, rely on AAC speech devices.” By acting as a “real‑time, two‑way speech and listening interpreter,” Pebble aims to reduce the social isolation that often accompanies reliance on manual translation during group work, freeing teachers to focus on instruction rather than transcription.
Honorable Mention: Utah TRAIL Lab
The final honorable‑mention project, Utah TRAIL Lab, shifts focus from generating new tutoring bots to evaluating the pedagogical quality of existing AI tutor responses. As described, “Given a student‑tutor conversation and a candidate tutor response, the goal is to assess whether the response appropriately recognizes the student’s mistake and provides useful guidance.” This quality‑assurance approach underscores a growing recognition that the effectiveness of AI in education hinges not only on generating content but also on ensuring that the content is instructionally sound, context‑aware, and aligned with proven teaching strategies.
Impact and Future Directions
Collectively, the hackathon’s outputs illustrate a vibrant ecosystem where AI is harnessed to enhance accessibility, equity, and instructional efficiency across diverse educational contexts. The prize structure—combining cash awards with AWS credits—provided teams with both immediate recognition and tangible resources to further develop their prototypes. Comments from the judging panel, though not quoted directly in the source material, emphasized the importance of keeping AI “as a supportive tool rather than a decision‑making authority,” a sentiment echoed in several project designs that maintain human oversight.
Looking ahead, the One‑U Responsible AI Initiative plans to build on this momentum by offering follow‑up mentorship pathways, potential pilot implementations in partner schools, and continued opportunities for interdisciplinary collaboration. As the field of AI in education matures, events like Minds and Machines serve as critical proving grounds where innovative ideas can be tested, refined, and ultimately scaled to benefit learners worldwide.
All quoted passages are drawn directly from the original event description provided by the One‑U Responsible AI Initiative.
Students build AI-powered solutions for education at Minds and Machines Hackathon