How Tenable Brought Multi-Language Support to Security Detectors With Agentic Translation
Tenable, Wednesday, October 21st, 2026
Tenable explains how it trained AI security detectors on nine more languages without a translation model in production.
Tenable explains how it made its AI security detectors natively multilingual, since prompt attacks in other languages are as dangerous as English ones.
It first used a translate-then-detect process but found it added cost and latency.
Tenable instead trained detectors on high-quality multilingual data covering nine additional languages.
The company says the key was keeping data labels intact, because translation can flip a malicious sample to benign or soften it.