Multimodal AI
Speaker verification
Checking whether a voice sample matches a claimed speaker identity.
Example
A service compares a caller's voice with an enrolled voice reference.
Why people use it
It provides one possible check that a caller matches a claimed identity.
What you'll hear
“Does the caller sound like the enrolled account holder?”
What this means for you
Use appropriate additional checks for consequential access decisions.
Can you control it?
Sometimes
Sometimes. Your choices depend on the tool and your access. The settings available to an everyday user may differ from those available to the people running it.
Common questions
- Can voice checking be fooled by imitation or generated audio?
- It can be vulnerable; resistance depends on the system and attack conditions.
- Is it the same as checking what the person says?
- No. It compares voice characteristics rather than whether the spoken statement is true.
- Can a genuine caller be rejected?
- Yes. Illness, poor audio or a changed speaking style can affect the match.