Privacy Architecture for Data-Driven Innovation Nishant Bhajaria - PowerPoint PPT Presentation

Privacy Architecture for Data-Driven Innovation Nishant Bhajaria

What is privacy? Unlike Security, privacy can be hard to define.

→ →

Confidential Intro - Nishant Bhajaria Staff Privacy Architect History: Nike ● Netflix ● Google Cloud ● Uber ● Mandate: Cross-functional technical privacy strategy

Privacy The Rules are changing

.

.

● ●

So what does this mean? ● Privacy is “all hands on deck” not just legal ● Security ≠ Privacy ○ Security is necessary but not sufficient for privacy ● Think beyond breaches ○ Data collection and Internal misuse ○ Data sharing and External misuse

Confidential Data Classification ● Answers questions ○ “What is this data?” ○ “How sensitive is this data?” ● Tiered ranking of user and business data

Data Classification Examples Data Example Example Data Classification Category Sets Tier 1: Highly Restricted Government Identifiers and location Social Security Card Driver’s License data (excludes personal data) License Plate Number Tier 2: Restricted Vehicle Data Proof of Insurance Make and Model Tier 3: Confidential Non-Identifying Vehicle Data Color Press Releases Tier 4: Public Public Information Product Brochures

Data Handling Collection Requirements “How can I protect Access this data?” Retention, Deletion, Sharing (internal/external)

Why is Data Inventory vital? Cannot apply data protection post collection without inventory Data Inventory External Collection Data Use Deletion and Tagging Sharing ● User Apps ● User Apps ● Retention Policy ● Web Site ● Export/DSAR ● Third-Parties ● Third Party Sharing

Data Sources Scanners/Classifiers UMS (In Metadata Manual -house global Data discovery (UI, Scanning and Decider metadata Inventory Crawlers, APIs,) detection store) DB (also supports AI models) UMS (In -house global metadata store) Other data sources ML-powered (Hive, classifiers Vertica, (automated MySQL, etc) data Deletion, detection) Retention and other privacy services

Data Sources Scanners/Classifiers Metadata Manual discovery (UI, Scanning and Crawlers, APIs, Data Decider detection Inventory etc) (also supports DB AI models) Other data sources ML-powered (Hive, classifiers Vertica, (automated MySQL, etc) data detection) Deletion, Retention and other privacy services

Metadata Sources UMS

Metadata Registry/Definition

Metadata Collection Pull model Push model ○ Crawler (periodic) ○ Automated e.g. sample data, stats e.g. data retention policies ○ Event-based (Event Listeners) ○ Crowdsource e.g. data quality e.g. table descriptions

• • •

• • •

• • •

• •

• •

•

• •

• •

• •

•

•

•

• ⇒

•