The source gives a measured accuracy comparison against single listeners, which helps readers gauge how far audio question answering can replace human judgment.
Hume AI has launched a new Hume Decision API, an audio-native model that can answer questions about a recording by analyzing the audio itself.
Users can write their own questions and answer options, such as whether a caller sounds frustrated, and receive an estimate of how human listeners would answer.
In initial tests, the model agreed with the panel's average answer 70.7% of the time, compared to 56.5% for a single listener.
Introducing the Hume Audio Decision API. Ask anything: How many speakers? Is anyone laughing? Is the caller frustrated? What used to take 10+ models is now one call. Trained on 10M+ human judgments, so it answers the way human listeners would. Realtime, or across hours of recordings. Request access below.
🚨 AI News | TestingCatalog · x.comPublished · added here
Original title: Hume AI has launched a new Hume Decision API, an audio-native model that can answer questions about a recording by analyzing the audio it...
