Operate an instrument and analyse a defined measurement
See the requirement-level assessment and its sources.
SCIENCE UNDER AUTOMATION
AI in Experimental physics: capabilities and evidence
Experiments require working instruments, calibrated measurements and a physical interpretation that survives systematic checks.
Autonomy · up to 4.4 / 10
Model interval: 2.1–4.4. Assessed 21 October 2026.
Two equally weighted workflow families: a bounded operational task and an investigation or deployment beyond that task. Inherited subfields use this parent scenario.
Calculation and work mix
Each requirement uses the operational scale from 0 (independent execution unestablished) to 4 (reliable independent execution in the stated setting). A workflow takes its lowest requirement level. The two workflows receive equal weight; the headline is 1 + 9 × weighted level / 4.
These levels are source-linked model assignments, not measurements reported by the papers. “Up to” is the upper model bound, not a probability or a forecast date.
Operate an instrument and analyse a defined measurement: 1–2 / 4.
Develop and verify a new physical measurement: 0–1 / 4.
Upper bound with 75:25 workflow weights: 4.9. With 25:75 weights: 3.8.
What has AI demonstrated?
An agent aligned samples on a real synchrotron beamline, identified reflections and recovered from an unexpected motor offset.
- Test
- Peer-reviewed physical instrument demonstration
- Comparison
- Comparators vary by study; see the linked sources.
- System and version
- An agentic artificially intelligent X-ray scientist · 2026-07-01
- Evidence checked
- 2026-10-21
The diffractometer, tool interfaces and virtual preparation were supplied. Alignment is one component of a scattering investigation.
What does the field require?
Operate an instrument and analyse a defined measurement
- Align and control prepared equipment
Autonomy level: 2–3 / 4.
A real beamline demonstration includes feedback and recovery from offsets, using prepared interfaces.
An agentic artificially intelligent X-ray scientist - Reconstruct the analysis and quantify outputs
Autonomy level: 1–2 / 4.
Collider-Bench reports uneven reconstruction and a continuing advantage for the physicist-in-the-loop solution.
Collider-Bench - Check calibration and systematic alternatives
Autonomy level: 1–2 / 4.
Alignment and reconstruction cover bounded checks, with expert experimental definitions supplied.
An agentic artificially intelligent X-ray scientist · Collider-Bench
Develop and verify a new physical measurement
- Specify the apparatus and discriminating experiment
Autonomy level: 0–1 / 4.
Prepared apparatus and experimental goals are inputs in the demonstrations; specialist design remains external.
An agentic artificially intelligent X-ray scientist - Build, commission and maintain the apparatus
Autonomy level: 0–1 / 4.
The tested agent operates an existing beamline rather than fabricating or commissioning one.
An agentic artificially intelligent X-ray scientist - Establish a new physical interpretation
Autonomy level: 1–2 / 4.
Reconstructed analysis supports interpretation, while incomplete implementation details and systematic uncertainty still require physicists.
Collider-Bench
Research tasks
Develop and verify a new physical measurement
See the requirement-level assessment and its sources.
Can AI replace the whole workflow?
The results above measure specific tasks. The field assessment also requires completion of the connected workflow, including the requirements listed here. A field-wide replacement percentage has not been established.
Read the assessment method