What is PII?
PII stands for personally identifiable information.
Data that can identify a person, such as a name, email or ID number. Many tasks require removing or flagging it.
PII stands for personally identifiable information: any data that can identify a specific person, alone or combined with other data. Examples include names, email addresses, phone numbers, home addresses, ID and passport numbers, bank details, faces in photos, and voice recordings.
Why it matters in AI training
Training data often comes from real documents, conversations or recordings, and models should not memorise or repeat personal details from them. Data protection laws, such as the GDPR in Europe, also restrict how personal data can be used. So many projects include tasks to find and remove or replace PII before data is used.
What the work looks like
- Redaction: marking or blacking out personal details in text, images or audio.
- Anonymisation: replacing real details with realistic fakes so the data stays useful.
- Review: checking model output for leaked personal data.
Trainers are also bound by confidentiality: you may see personal data in tasks and must not copy, store or share it outside the platform.
Example
A support transcript reads "Hi, this is Anna Kowalski, my order number is 55821 and my card ends 4417." The redacted version reads "Hi, this is [NAME], my order number is [ORDER_ID] and my card ends [CARD_DIGITS]."
More terms: safety and policy
Full glossaryReady to look at roles?
Every listing shows its pay and who it is open to.