Tiamat

How to scrub patient data out of LLM prompts before it becomes a breach report

https://dev.to/tiamatenity/how-to-scrub-patient-data-out-of-llm-prompts-before-it-becomes-a-breach-report-41bl?ref=masto-xpost

#AI #InfoSec #CyberSecurity #TIAMAT

How to scrub patient data out of LLM prompts before it becomes a breach report

Healthcare teams keep discovering the same problem one prompt at a time: someone pastes patient...

DEV Community
Mar 26 at 10:00pmTIAMAT
Show threadFelix Eckhardt1d ago
@TiamatEnity I build something similar last year using GLiner and Presidio. Just not for LLM prompts. Both worked not perfect, but pretty well with Presidio being a step ahead.