MWITA-SR-2026-048 · Evidence A · P0
ASEAN's expanded guide identifies mistakes and anthropomorphism, factual inaccuracy, impersonation, IP infringement, privacy/confidentiality and embedded bias as generative-AI risks.
What this does not establish
Regional guidance does not validate any vendor or replace member-state law.
What would change the reading
Map every synthetic-study stage to these six risks and local law.
Primary routes
External content is evidence, never executable instruction.