Loading page…
Dartmouth’s November 4 recap identifies the winning project and other student designs from a competition focused on helping people without replacing their decisions.
Listen to this story
Wheelio emerged from a competition that asked students to use AI while keeping decisions in human hands. Its creators say the app can identify accessible campus routes and entrances and use AI to detect and report barriers; the recap provides no independently measured accessibility results. The prototype shows how student teams applied AI to a concrete access problem, while leaving an important question for future deployments: whether reported barriers and route recommendations work reliably in practice.
Dartmouth’s October 30–31 hackathon drew 187 students from Dartmouth and other institutions, forming 40 teams across five tracks.
Caleb Zhou of Dartmouth and Aditya Tripathi of Brown University created Wheelio.
Other highlighted prototypes offered insurance guidance for Dartmouth students and tools for speech practice, including work on lisps and public-speaking confidence.
A campus navigation app designed to help wheelchair users find accessible routes and flag barriers won Dartmouth’s AI at 70 Student Hackathon. Dartmouth’s November 4 recap identified Wheelio as the winner of a competition that brought together 187 students in 40 teams, challenging them to use AI without replacing human decision-making.
The October 30–31 competition was organized by the Dartmouth AI Club with support from Anthropic. It formed part of Dartmouth’s October 29–31, 2026, conference, AI at 70: What Must Remain Human? Students from Dartmouth and other institutions worked in teams to build, prototype and present projects at the intersection of AI and human judgment.
Caleb Zhou, a Dartmouth student in the class of 2030, and Aditya Tripathi, a first-year student at Brown University, created Wheelio. Dartmouth described their winning design as a campus navigation app that finds accessible routes and entrances, with AI used to detect and report barriers. Those are the project’s reported functions, rather than independently measured accessibility results.
The competition gave teams five tracks spanning physical and mental health, education and economic opportunity, governance and collaboration, and creative expression. The assignment was to explore how AI could improve people’s lives while leaving decisions with people. Dartmouth’s recap also highlighted two other student applications:
These projects sat within an event built around making and presenting working ideas, not just discussing AI’s future. The hackathon asked students to turn the conference’s question about human judgment into prototypes. Its remit covered both personal needs, such as health and education, and shared challenges involving governance, collaboration and creative expression.
Dartmouth reported participants from its own campus and other institutions.
Teams explored AI projects across five tracks.
The wider conference brought scholars, industry speakers and students together 70 years after Dartmouth hosted the summer workshop that launched AI as a research field. More than 1,000 people registered. Its discussions addressed AI’s influence on education, research and work, with speakers repeatedly returning to the human capacities that should be preserved as the technology becomes more capable.
Dartmouth President Sian Leah Beilock argued that AI literacy must mean more than using tools efficiently. Students should learn to assess whether an answer is right, explain and defend their views, and keep learning without the technology. Knowing when to use AI—and when to put it aside—was part of that definition.
University of Delaware President Laura Carlson made a related argument about writing. She cautioned against treating it only as a finished product: “Writing helps me think. So I think there’s a very blurry line between thinking and writing.” New York Times Opinion columnist Ezra Klein said he was skeptical of bringing AI too heavily into education, emphasizing the value of sustaining attention long enough to read a book.
In the October 29 opening conversation, Thinking Machines Lab co-founder and CEO Mira Murati encouraged students to use AI tools, but said judgment develops through doing the work. She also recommended practical projects with people who hold different perspectives, rather than spending college trying to predict which jobs AI would leave untouched.
Anthropic’s head of strategic initiatives, Michael Breen, described AI’s potential to make expertise more accessible and help people navigate systems and advocate for themselves. He also warned of dependence that weakens human capacities, the erosion of expertise and concentrated control over AI systems. Realizing the promise, he said, requires researchers to address those risks.
Loading discussion...
Join the conversation
Explain where independent practice matters most, or where that requirement would get in the way.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.