Research Summary: ChatGPT Solves All Tested Qiskit Homework Assignments

Generative AI, specifically ChatGPT, successfully completed all tested introductory Qiskit homework assignments, highlighting a significant challenge for assessment methodologies in quantum software education. The research explored strategies to adapt homework design to require student engagement beyond simply submitting AI-generated code, without outright banning AI tools.

Why it matters

The demonstrated capability of generative AI to complete technical assignments poses a critical strategic challenge for educational institutions and professional development programs across all technical domains. Organizations must consider how to adapt learning and assessment strategies to prepare individuals for a future where AI tools are integral to work processes, ensuring that education fosters critical thinking and application rather than mere content generation.

Key insights

• ChatGPT is capable of solving introductory Qiskit homework assignments, including tasks involving seeded basis-state circuits, Quantum Fourier Transform, and Deutsch-Jozsa algorithms.
• The study aimed to develop assessment methods that leverage rather than prohibit generative AI by requiring students to execute, review, and discuss results.
• Proposed homework designs incorporated personalization, simulator execution, JSON submissions, hidden references, circuit metrics, reflections, and optional IBM Quantum execution.
• The findings suggest that current assessment paradigms in quantum software education may need re-evaluation given the capabilities of advanced AI models.

Source

arXiv — Computers and Society — https://arxiv.org/abs/2608.19707