Leverage AI to Protect Your Data Assets... from AI
summary
Claim: flexible NoSQL data hides PII and AI-driven code changes speed leaks unless data is governed. Demonstrates schema extraction from live MongoDB/JSON blobs (one doc had ~14,000 lines later refactored into ~20 entities), automated PII scanning (heuristics plus a BERT NLP prototype that ran ~3 minutes on CPU and needs GPUs), and a streaming pipeline (Kafka/Pub-Sub, UDF transforms) that enforces deterministic masking pipelines via Studio 3T/MCP and 3T Bridge so downstream consumers only see anonymized events.
AI-written summary · about
chapters (9)
- 0:00Data leaks in AI development
- 2:12PII risks in NoSQL schemas
- 5:24Vector embeddings leaks
- 7:49Data governance gaps
- 9:13Mitigating data leaks
- 15:58Schema extraction and PII scan
- 21:31Automated data masking
- 25:41AI and data protection needs
- 26:23Q&A
related talks
25mSafeguarding Sensitive Data in the Age of Generative AI
40mEnsuring Data Security in the AI Revolution
46mProtecting sensitive data in huge datasets: Cloud tools you can use
51mCloud data and AI security in 2024: What you need to know
42mRise of the Machines: Achieving Data Security and Analytics with AI
39mData Privacy Engineering to power Responsible Enterprise AI
40mRolling out Enterprise AI: Tools, Insights, & Team Empowerment
34mInvisible Infiltration of AI Supply Chains: Protective Measures from Adversari...