Interactive teaching tool · Responsible AI · Human review

AI Use Decision Lab

Decide what role AI should play before you prompt it. Classify the use, surface privacy and consequence risks, define human verification, and draft a transparent record or disclosure note.

Local-first · nothing entered here is sent to an AI system

1. Define the use

What are you trying to accomplish?

Planned AI role

Context

Applicable policy or rules

2. Surface the risk signals

Personal, confidential, proprietary, student, employee, client, or unpublished data
Current facts, laws, policies, prices, schedules, or recent research matter
Quotations, citations, links, references, or source attribution matter
Calculations, code, data transformations, classifications, or quantitative conclusions matter
An error could materially affect health, employment, education, finance, eligibility, safety, rights, or access
The use asks AI to infer a person's traits, motives, emotions, intent, identity, or risk
The result will be submitted, published, sent externally, graded, or used in a formal decision
The system would act, score, send, publish, or decide without explicit human approval

3. Make the record transparent

What did AI actually contribute?

What did the human author/reviewer do?

What was independently checked?

AI use record / disclosure draft

Describe the contribution and human review to build a draft.

This is a transparency aid, not a universal disclosure rule. Course, employer, publisher, funder, client, and jurisdictional policies may require different language—or prohibit the use entirely.

Disclosure is not citation

Disclosure explains how AI was used in the work process.

Citation/attribution identifies evidence, sources, or quoted ideas.

Verification checks whether material claims are correct.

One does not substitute for the others.

Teaching prompts

Brainstorming lesson examples

What changes if AI generates possibilities but the learner must justify the final choice? What evidence of thinking should remain visible?

Summarizing a supplied article

How do source-bounded instructions, quotation checks, and author-vs-model interpretation change the review plan?

Feedback on student work

What student data may be shared? Who owns the evaluative judgment? What appeal or correction process exists?

Hiring or eligibility decision

Which parts, if any, are appropriate for AI assistance? Which judgments must remain with accountable humans? What inferences should not be made?

Method note. This tool extends the same teaching principle as Prompt Lab: better prompting does not eliminate the need for evidence, privacy controls, human judgment, and review proportional to consequence. It is designed to make those decisions inspectable before AI use. It is an educational planning aid, not legal, compliance, institutional-policy, or professional advice.