Caribbean AI Association
Back to the archive
Research papers

LAC Eval: testing AI services for local use

Adrian Dunkley   |   2026-11-10 · Research edition 0.1 · evaluation protocol

A development set contains 36 cases across three service sectors and six risk categories. A scoring guide records task completion, use of evidence, action limits and referral. No model performance results are reported.

Archive print edition of the original web text

Read the original publication