PII (Personal Identifiable Information)
Information that can identify a specific person alone or in combination, such as name, phone, ID number and address.
In plain language
In AI systems, PII risks appear in training data, prompts, logs, retrieval stores and model output. Removing names is not enough: birth date, postcode and job title can re-identify someone. Engineering controls include data classification, minimization, anonymization/pseudonymization, access control, retention limits and auditing. Before calling a third-party model, confirm whether data is used for training, where it is stored and for how long. Cross-border transfers add compliance requirements.
Example
A support log saying “Ms. Wang, 138****, lives in a Chaoyang community” may still identify someone even with the phone number masked.
Recommended tools
Related terms
Alignment
Research and engineering to make AI systems behave according to human intent, values and safety constraints.
Guardrails
Rules and systems that constrain model inputs, outputs and actions.
Retrieval-Augmented Generation
Retrieve relevant passages from a knowledge base, then have the model answer from them.
Agent Memory
A mechanism for an agent to store and retrieve user preferences, past tasks and key facts across sessions.