Pilot for AI-assisted walking challenge planning

Inputs for the first review

• The purpose of the program and intended participants
• Timing and the people who will operate it
• The requested behavior and a reasonable upper limit
• The budget approach and the person authorized to approve spending
• Available data and its permitted use
• Legal, security, and organizational constraints

Pilot deliverables

• A draft challenge: participants, period, goal, and participation conditions
• A draft participant explanation and consent checklist
• A responsibility map for AI, people, and ESPL
• A list of data fields and evaluation measures
• Legal, security, and operational questions to resolve before implementation

AI-generated text is not published as-is. A person checks facts, feasibility, accessibility, participant burden, and potentially discriminatory effects.

If the proposal includes money

AI may suggest a budget or distribution structure. The sponsor or authorized person verifies the amount, recipient, conditions, network, fees, and legal and accounting treatment, then initiates any wallet action. This human-approval policy is the same in the Japanese and English descriptions.

Not included

• Development or operation of a production AI agent
• Smart-contract implementation or audit
• Procurement, custody, or transfer of JPYC or other assets
• Final legal, tax, medical, or security decisions
• Guarantees of participation, adherence, health outcomes, or commercial results

Process

1. Confirm the purpose, participants, and constraints.
2. Generate several drafts and have the owner narrow them down.
3. Review burden, data, funding, legal, and operational risks.
4. If appropriate, test with a limited group and period.
5. Have people decide whether to continue, revise, or stop.