Beatriz Yankelevich of MIT’s Engineering Quantum Systems Group used GPT‑5.6 Sol, integrated with Codex, to autonomously run and refine routine measurements on superconducting qubits. This integration allows the AI agent to operate laboratory software, analyze results, and decide on subsequent steps without constant human supervision.
- The system successfully completed standard measurement sequences on an uncalibrated six-qubit chip, identifying transition frequencies and determining quantum information retention times.
- It handled clear experimental signals effectively but struggled with weak or noisy data, requiring researcher guidance to find suitable parameters.
- The automation frees researchers from monitoring calibration processes, allowing them to focus on experiment design and data analysis while agents run overnight.
The approach demonstrates that AI agents can handle well-defined experimental workflows, significantly reducing the time required for routine chip characterization.