Pair AI-permitted take-home assignments with supervised modifications, oral defense, prediction, and transfer tasks to verify understanding¶

Strategy · All strategies
Evidence · no claims cited

Description¶

A concrete response to LLM outsourcing: allow students to use AI for syntax, setup, or debugging while instructors directly assess whether they can explain and adapt the result. Practical options named by the article include a short oral defense tied to the student's own circuit and counts, a supervised modification of a qubit mapping, oracle, or input state, prediction of a new outcome before execution, and a transfer task that changes the algorithmic structure rather than only the seed. The article lists these as recommendations derived from its observed failure mode.

Design Implications¶

Context¶

Requirements¶

• Instructors must be able to interact with students around their own submitted artifacts

Constraints¶

• These measures were not experimentally compared in this study

Target Learners¶

• undergraduate quantum-computing and programming students

Target Learning Goals¶

• explaining and adapting one's own code
• predicting circuit outcomes before execution
• transferring algorithmic understanding across structural changes

Related Strategies¶

• Use weekly oral code review interviews to incentivize students to understand AI-assisted code
• Replace detector-centred assessment with a four-stage process-oriented model: declared starting point, versioned drafting, process artefacts, and targeted oral verification
• Scaffolded critique-of-AI-output assignment sequence with paired verification of individual understanding
• Redesign assessment toward supervised, oral, process-based and authentic formats to verify students' own reasoning under GenAI
• Stage GAI use in assessment: permit exploration, restrict independent reasoning, require justification
• Mini homework defenses: short oral one-on-one assessments of student solutions
• Two-instrument assessment: AI-free coding quizzes plus verification-gated oral defenses of comment-stripped AI-assisted work
• Provide AI-free proctored spaces so some take-home work is verifiably unassisted

Examples¶

-

Key Sources¶

• Alexei Kaltchenko and Gurnivaj Tiwana. (2026). ChatGPT Solves All Tested Qiskit Homework Assignments. arXiv. https://arxiv.org/abs/2608.19707