AI Translation Analysis
Resources and analysis on AI translation quality.
Articles and insights
These articles examine where AI translation actually fails in regulated sectors — and why it matters. The analysis draws on forty years of professional translation in maritime, legal, financial, and pharmaceutical contexts, including senior roles at NTTData and Accenture, and applies that specialist knowledge to the specific failure modes that machine translation produces when it encounters thin training data and compressed domain expertise. This isn't a case against AI translation. It's a case for knowing where the risk sits.
AI Translation Works for Hotels. Pharmaceutical, Legal and Financial Documents Are Different.
Enterprise AI translation delivers proven results for high-volume content. The risks are different in pharmaceutical, legal, maritime and financial translation — and the error tolerance is not the same.
The Hidden Cost of AI Translation Without Post-Editing
AI translation without post-editing produces fluent text — and errors in meaning, register and terminology that no engine detects alone.
When AI Gets It Wrong and Nobody Checks: Lessons from Europe's Disinformation Crisis
The same structural failures documented in Europe's disinformation crisis — overconfident automated systems, failed crowdsourced verification, and the erosion of expert oversight — are present in AI translation in regulated sectors.