Notesfoundations🌱 seedling
Data Ethics
Also: Responsible Data Use · Doing the Right Thing
Data ethics is the practice of deciding what data about people a system collects, keeps, infers and acts on, and who could be harmed by it. Most of those decisions are made by engineers in schemas, logs, retention settings and training sets, often without noticing they are decisions.
Why it exists
Data about people can be breached (permanently), repurposed beyond what they agreed to, subpoenaed, sold (even in a bankruptcy), and used for automated decisions about jobs, loans and benefits. A model trained on it can learn and scale the biases in the decisions that produced it.
How it works
Treat personal data as a liability with a use. The GDPR's principles work as design rules anywhere:
- Lawful, fair, transparent: people can tell what happens to their data.
- Purpose limitation: collected for a specified purpose, not reused for unrelated ones.
- Data Minimization: adequate, relevant and limited to what is necessary.
- Accuracy, and storage limitation (Data Retention).
- Integrity and confidentiality, plus accountability: be able to demonstrate all of the above.
- Rights that become engineering requirements: access, correction, erasure, and limits on purely automated decisions.
Trade-offs
| Gains | Costs |
|---|---|
| smaller breach and legal exposure | some analytics and features become impossible |
| user trust | engineering effort: deletion pipelines, audits |
| fewer surprises from regulators | slower "collect everything" experimentation |
In practice
- Fines are now design-level: Ireland's DPC fined Meta €251M in 2024, €240M of it for failing data protection by design and by default (Privacy by Design).
- The ACM Code of Ethics (1.6) asks that "only the minimum amount of personal information necessary" be collected.
Pitfalls
- Assuming legal review will catch it: by the time a lawyer sees the system, the data already exists.
- Treating ethics as only about privacy; Algorithmic Bias harms people without any data leaking.
Connections
- Part of
- Foundations
- Related
- Privacy by DesignReliability
- Mentioned
- Data Minimization
Recall
These cards are also in your review deck, scheduled with spaced repetition.