Ethics · Established · Beginner

Data Privacy in AI

Protecting personal information across the AI lifecycle — collection, training, prompting, logging and output.

What Data Privacy in AI is

AI creates new privacy surfaces: prompts containing personal data, logs retained by vendors, models that can regurgitate training text, and retrieval indexes that ignore access controls.

How it works

Controls include minimising personal data in prompts, redaction before sending, contractual limits on vendor training and retention, per-user permission filtering at retrieval, short log retention and honouring deletion requests through the whole chain.

Why it matters

Privacy failures in AI features are among the most common and most expensive incidents, and they undermine trust immediately.

Common uses

→Vendor and DPA review
→PII redaction pipelines
→Retention policy for prompt logs
→Access-aware retrieval

Strengths

✓Enables use of AI on sensitive workloads when done properly

Watch for

✓Adds friction
✓Regurgitation risk is hard to fully eliminate

Continue exploring

AI Data Exfiltration Edge AI AI Governance AI Regulation