Sierra Unveils fleming-1 AI Voice Detection Model for Calls
Sierra's fleming-1 scores callers' voices in real time to flag AI-generated audio, with a conservative default to protect real humans.
Sierra has announced fleming-1, a model designed to detect whether a caller is a human or an AI agent, according to digitaltoday.co.kr.
fleming-1 analyzes the caller's voice in real time during a call and assigns a score indicating whether the voice is AI-generated. Calls judged to be AI are marked, and the default setting is conservative to avoid misclassifying real humans as AI.
The model only provides information and does not make decisions; the company using it decides how to respond. fleming-1 can be used in all voice agents built with Sierra.
In October 2025, Sierra announced the Personal Agent Protocol in cooperation with other tech companies. According to the company, distinguishing human and AI voices on calls is becoming increasingly difficult as AI voice quality improves, with noise or unstable connections further hindering judgment.
Quick answers
What is fleming-1?
fleming-1 is a model from Sierra that detects whether a caller is a human or an AI agent by analyzing the caller's voice in real time and assigning a score.
Does fleming-1 decide how to handle a call?
No. fleming-1 only provides information; the company decides how to respond.
Where can fleming-1 be used?
It can be used in all voice agents built with Sierra.