Notesfoundations🌱 seedling

Privacy by Design

Also: Data Protection by Design · Privacy by Default · Data Protection by Default

Privacy by design means building data protection into a system's architecture from the first design rather than bolting it on later; privacy by default means the out-of-the-box setting processes only the data each purpose needs, with sharing opt-in.

Why it exists

Retrofitting privacy is expensive and incomplete: once data has spread into logs, warehouses and partners, it can't be un-collected. Decisions made in the first schema and the first defaults dominate everything after.

How it works

  • Seven foundational principles (Cavoukian, 2009): proactive not reactive; privacy as the default setting; embedded into design; full functionality (positive-sum); end-to-end lifecycle protection; visibility and transparency; respect for the user.
  • GDPR Article 25 makes it law. By design: technical measures such as pseudonymisation and Data Minimization chosen when the system is designed. By default: only data necessary for each specific purpose is processed, covering amount collected, extent of processing, storage period and accessibility.
  • In engineering terms: privacy review at design time, data inventories, defaults that don't share, Data Retention limits and a deletion path planned before launch.

Trade-offs

Gains Costs
cheaper than retrofitting design-time effort and review
smaller blast radius when things go wrong fewer "free" data for future ideas
defensible under regulation opt-in sharing lowers engagement with optional features

In practice

  • Meta, Dec 2024: Ireland's DPC fined it €251M over the 2018 "View As" token breach; €240M was for violating Article 25 (by design €130M, by default €110M).

Pitfalls

  • Dark patterns: technically opt-in defaults wrapped in UX that pushes everyone to opt out.
  • A privacy review after the data model is fixed: it can only add paperwork.

Recall

These cards are also in your review deck, scheduled with spaced repetition.