Security & Privacy
Membership inference
An attack that tries to determine whether a specific record was in a model's training data.
Example
An attacker tests whether a particular person's record appeared in the training data.
Why it matters
It helps assess privacy exposure even when an entire original record cannot be recovered.
You'll hear it
“Could someone tell that this person's record was used for learning?”
Can you control it?
Set by the team
The people building or running the AI choose this setup. An everyday user generally needs their help to change how this part works.
Common questions
- Does a successful guess reveal the full record?
- No. It may reveal only that the record was likely included.
- Are all guesses certain?
- No. An attack can make false assumptions and needs careful evaluation.
- Can inclusion alone be sensitive?
- Yes. Belonging to a particular collection can reveal private circumstances.