Design and simulate an artifact against a specified brief
See the requirement-level assessment and its sources.
SCIENCE UNDER AUTOMATION
AI in Engineering: capabilities and evidence
Engineering must turn requirements into an artifact that performs under physical constraints, including failure conditions.
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.
Design and simulate an artifact against a specified brief: 1–2 / 4.
Commission and verify a new physical system: 0–1 / 4.
Upper bound with 75:25 workflow weights: 4.9. With 25:75 weights: 3.8.
What has AI demonstrated?
Agents generate and revise CAD using explicit engineering tools and a supplied load case; the evaluation reports increased structural complexity and compile rate.
- Test
- Peer-reviewed IJCAI engineering benchmark
- Comparison
- Comparators vary by study; see the linked sources.
- System and version
- Physics-in-the-Loop · 2026
- Evidence checked
- 2026-10-21
Computational verification is within the supplied model. Fabrication, commissioning and physical certification are outside the demonstrated workflow.
What does the field require?
Design and simulate an artifact against a specified brief
- Translate requirements into a coherent design
Autonomy level: 1–2 / 4.
CAD benchmarks show progress with tool feedback and substantial failures on complete typed requirements.
Physics-in-the-Loop · CAD agents with finite-element feedback, v2 - Generate and revise executable models
Autonomy level: 2–3 / 4.
Agent loops generate designs and revise them using computational feedback.
Physics-in-the-Loop - Check physical constraints and failure criteria
Autonomy level: 1–2 / 4.
FEA makes structural checks possible; no strict-passing first-attempt artifact appeared in the cited harder sweep.
CAD agents with finite-element feedback, v2
Commission and verify a new physical system
- Fabricate and integrate the complete system
Autonomy level: 0–1 / 4.
The cited design evaluations end with computational artifacts; real integration remains specialist work.
Physics-in-the-Loop · CAD agents with finite-element feedback, v2 - Measure performance under real operating conditions
Autonomy level: 0–1 / 4.
Simulation validation does not perform physical commissioning or field testing.
Physics-in-the-Loop - Resolve failures and accept the deployed system
Autonomy level: 1–2 / 4.
Automated feedback helps iterate designs, while responsibility for physical validation remains outside the demonstrated agents.
CAD agents with finite-element feedback, v2
Research tasks
Commission and verify a new physical system
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