Secure AI: De-identifying data with SDP

Опубликовано 2 октября 2025, 23:09

Learn how Google Cloud’s Sensitive Data Protection (SDP) addresses the challenge of using realistic data for development and testing without introducing security risks, especially for generative AI applications. This video explains SDP's detection and transformation capabilities, including redaction, masking, and tokenization. Discover how SDP secures user prompts, application logs, and model outputs, and learn about Model Armor for real-time generative AI security.

Chapters:
0:00 - Intro
0:25 - What is Sensitive Data Protection (SDP)
0:36 - SDP's Detection Capabilities
1:29 - De-identifying Data in Cloud Storage (Demo)
2:35 - SDP for Generative AI Prompts
3:22 - Model Armor for Real-time AI Security (Demo)
4:04 - Summary

Resources:
SDP Documentation → goo.gle/46C5pVb
SDP Cloud Skill Boost → goo.gle/3W8Ot3S
Model Armor Cloud Skill Boost → goo.gle/4mDR1C7

Subscribe to Google Cloud Tech → goo.gle/GoogleCloudTech

#GoogleCloud #SecureAI #SDP #ModelArmor

Speakers: Aron Eidelman
Products Mentioned: Google Cloud Security