Privacy Review for Non-Maniacs
summary
Meta reworks privacy review from manual checklists into upfront requirements plus automation, claiming ~500 privacy requirements and roughly half now enforced by automated verifiers and audits that produce stapled evidence. Engineering controls run on pull requests to comment on pull requests or block merges if deletion/retention checks fail; experiments use LLMs to detect consent flows when pattern matching is hard. Result: fewer ad‑hoc human decisions, faster consistent reviews and measurable coverage.
AI-written summary · about
chapters (7)
- 0:00The test review thought experiment
- 2:39Shift left and automation
- 3:37Privacy impact assessments
- 5:50Defining the compliance role
- 7:57Automating compliance outcomes
- 10:07When to use impact assessments
- 14:01Data and concluding remarks
related talks
12mA People-First Approach to Introducing Process and Tools for Technical Privacy Review
28mPrivacy Engineering Networking Break AMA
16mAutomating Technical Privacy Reviews Using LLMs
16mCache-22: Doing Privacy Engineering with Privacy Standards
26mTechnical Privacy Review Workshop
37mFrom the Trenches Privacy Engineering Pitfalls Successes
19m"A Big Deal": How Privacy Engineering Can Streamline the M&A Process
30mMaking Privacy Assessments More Efficient