Robotics evidence guide · automated editorial preparation, not expert review
Start assistive robotics with a bounded human need
Choose the assistance task with the people affected and define what useful support would mean to them. A humanoid demonstration is not enough to establish a beneficial service.
Evidence and interpretation
NIST’s risk framework gives broad motivation for considering affected individuals. Maha’s proposed task-selection worksheet asks what burden the robot is intended to reduce and what new burden it might introduce. Waiting, setup, supervision and recovery can matter as much as the successful motion. Keep the beneficiary’s preference separate from the operator’s convenience or the vendor’s capability claim. Record when an existing non-robotic alternative meets the need more simply; the worksheet is not designed to ensure that a robot is always selected.
Proposed evidence workflow
- Define one assistance outcome and the intended setting with affected participants.
- Compare the complete workflow, including setup and recovery, with available alternatives.
- Record excluded uses and unanswered suitability questions before planning any human-facing trial.
Worked illustration
A proposed tray-delivery service completes the transfer but requires a person to prepare every object and monitor every attempt. The evaluation should count that workload rather than equating motion completion with reduced burden.
Limits
No human study, clinical benefit or assistive suitability has been established. Physical trials require separate qualified review and safeguards.
Sources and review
Locator: Overview of the AI RMF
The framework is voluntary and addresses risk considerations across AI design, development, use and evaluation.
Boundary: This overview does not determine applicable law, certify robots or validate Maha workflows.
Inspection and reuse basis
Selected sections inspected 2026-09-15. Original explanation and link only; no framework text redistributed. This is not independent expert review of this article.