Season 2 candidate preview — not a launch or delivery commitment. Sample sizes, prices and timing are illustrative, pending qualified-panel measurement, safety tests, legal review and an approved contract.

Pilot proposal

Eval · Prompt Set

Lay-person evaluation and prompt data for health AI

A proposed health-chatbot comparison and general-question task. Collection and delivery are paused; grade A is not human verification, expert labeling or safety validation. Scope and output rights require separate approval.

Paid pilots also remain gated on legal review, eligibility and output-rights terms. Dedicated task UI is in development.

Who it is for

Model-evaluation teams at medical and health LLM / chatbot companies
Localisation teams at multilingual health apps in Southeast Asia

Example questions

“2,000 A/B comparisons of patient-facing chatbot answers
“3,000 realistic general health questions in Vietnamese, Thai and Bahasa (no personal details)

Specification

Contributors: Grade A only · members who accepted the AI-task terms (T3)
Quality: Generated-text detection, duplicate/personal-health exclusion and human review are unvalidated candidate procedures; service scope unconfirmed
Acceptance and rework: Sampling rate, rejection threshold and free-rework terms unapproved; no guarantee before contract

Included

Delivery paused; ratings, reason codes, written items and contributor IDs are not approved outputs. Random IDs do not authorize respondent-row or verbatim delivery

Not included

Any member health information
"Medical expert" labels — describing lay-person evaluation does not authorize current sales

Price table

Eval · Prompt Set	by task time	₩600 (KRW) / min
Minimum contract		₩3,000,000 (KRW)

Illustrative internal estimate · not an offer · sample, price, timing and payment terms unconfirmed