We don't have a scalable way to verify that AI-powered translation tools used in legal or medical settings are fully accurate.
open
Global / Unspecified, Global
WS00790
Mistakes in AI-assisted translation can have serious consequences in legal or medical contexts, yet independent verification of accuracy in these high-stakes settings remains inconsistent. Building reliable safeguards for critical-use translation remains an unresolved challenge.