Use this when an AI proposal sounds useful but its effects fall on different people. Look at the benefit, the possible harm, and what evidence would make the tradeoff clearer.
Prompt template
Analyze the proposed AI use: [scenario]. Identify who benefits, who bears risk, what can be measured, and what cannot. Consider privacy, access, error consequences, labor, and environmental costs where relevant. Suggest safeguards and remaining tradeoffs. Label uncertain claims. Boundary: Produce an analysis and proposed safeguards only. Keep the supplied scenario and facts intact. Leave policies and systems unchanged; don't make domain decisions or authorize deployment.
Fill in this prompt in the builder
What to supply
- The proposed use, affected people, and information involved.
- Known constraints and what remains uncertain, included in the scenario.
Worked example
Illustrative example: the input and result below are invented.
Input
Nora, 42, is a nurse educator considering drafts of plain-language versions of already-approved general training material. No patient information is involved. Educators would review every draft, and the original would remain available. No learning or time-saving results are supplied.
Filled-in prompt
Show the complete prompt
Analyze the proposed AI use: Nora, a nurse educator, is considering plain-language drafts of approved general training material. No patient information. Educators review each draft; originals remain available. Learning effects, review effort, account retention, and resource use are unknown. Identify who benefits, who bears risk, what can be measured, and what cannot. Consider privacy, access, error consequences, labor, and environmental costs where relevant. Suggest safeguards and remaining tradeoffs. Label uncertain claims. Boundary: Produce an analysis and proposed safeguards only. Keep the supplied scenario and facts intact. Leave policies and systems unchanged; don't make domain decisions or authorize deployment.
Result
Possible benefit: Learners may find unfamiliar terms easier to follow. Evidence needed: Can they explain the original instruction accurately? Risk bearer: Learners and educators if a simplification changes a qualification. Safeguard: Compare each material instruction with the approved original. Labor: Drafting could shift effort into review; no saving is established. Privacy and access: Keep patient information outside the process and confirm account handling. Environmental impact: No task-specific usage or energy evidence is supplied, so leave the size of that impact unresolved. Remaining tradeoff: Readability gains must be weighed against review effort and semantic accuracy.
What I’d check or correct
I'd check that a proposed readability benefit didn't become a claim that learning improved. I'd keep the example about general training material and leave clinical advice outside its scope.
What to check
- Benefits and harms identify the people affected and evidence needed.
- The analysis preserves the supplied scenario and doesn't redesign policy, make clinical decisions, or authorize deployment.
- Safeguards address the specific risk, with residual tradeoffs still visible.