evidoria

← Back to browse

Good practice Imported

Classmoji AI-Driven Oral Examinations for Code Assessment — Dartmouth College

United States of America · Hanover · See the United States of America profile · See the Hanover profile

Evidence: Descriptive / self-reported Top 63% 47/100 · Ask Evidence Copilot about this practice

Dartmouth's open-source Classmoji platform adds AI-driven oral exams that question students conversationally about their own submitted code, probing design choices and edge cases to check real understanding versus unexamined AI-generated work; presented at ACM SIGCSE TS 2026.

Classmoji AI-Driven Oral Examinations for Code Assessment — Dartmouth College

Details

Maturity
Pilot
Promoter
Dartmouth College — Classmoji (open-source GitHub-native CS education platform)
Period
2025–2026 (presented at ACM SIGCSE TS 2026, Feb 2026)
Keywords
higher education, computer science, academic integrity, open-source

Context

Classmoji is an open-source, GitHub-native learning-management platform built by educators and students, initially at Dartmouth College, treating GitHub organisations as classrooms and issues as assignments with an emoji-based feedback system. Dartmouth researchers added an AI-driven oral-examination feature that conversationally questions students about the code they submit, probing design choices and edge cases to distinguish authentic understanding from AI-generated code students cannot explain.

Activities

The approach is documented in the peer-reviewed paper 'AI-Driven Oral Examinations for Code Assessment: Evaluating Understanding Beyond the Commit', presented at the 57th ACM Technical Symposium on Computer Science Education (SIGCSE TS 2026, February 2026).

Conclusions

The evidence available is a peer-reviewed description of the system's design and rationale rather than published outcome statistics — no accuracy figures, grading-correlation data or student-sentiment scores were found; the practice's strength lies in open-source transparency and formal academic peer review, not yet in measured learning impact.

Implementation
Indicative cost
Low (< €50k)
Time to results
Short (< 1 year)
Staffing & skills
Dartmouth College researchers and educators (Classmoji project team), Student contributors to the open-source GitHub-native platform

Conditions for success

  • Built on an already-adopted open-source LMS (Classmoji) rather than a standalone tool
  • Adaptive follow-up questioning tied directly to each student's own submitted code
  • Public, peer-reviewed documentation of the methodology (ACM SIGCSE TS 2026)

Common failure modes

  • No published accuracy, grading-correlation or student-sentiment data yet
  • No evidence found of adoption or oral-exam feature use outside the originating Dartmouth team

Where it fits

Governance type
higher-education institution (single university)
Scale
single-institution pilot within an open-source platform
Income level
high-income

Commonly funded by

Own resources / municipal budget

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Replication kit

Reusable artefacts from this practice — as published by their sources.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful