In April 2024, New York-based Hume AI launched its Empathic Voice Interface (EVI), an emotionally intelligent voice interface that can be built into any application.
Unlike text-based chatbots like ChatGPT and Claude 3, Hume is primarily a voice-based interface – listening to variations in the human user’s voice, including intonation, pitch and pauses.
Trained on photos and audio data from hundreds of thousands of people around the world, according to Hume, the interface goes beyond the typical universal emotions such as ‘happy’, ‘sad’, ‘angry’ or ‘afraid’, instead providing a list of 53 different emotions on its website that it is capable of detecting from users. This helps to give the AI a more nuanced and multidimensional take on human emotions and give it a more granular understanding of user needs.
These are reflected in Hume’s “speech prosody model”, which “measures the tune, rhythm, and timbre of speech and is incorporated into EVI” and which convey up to “48 distinct dimensions of emotional meaning”. The model displays the proportion of different emotions detected in the right hand sidebar.
The company was founded by Dr Alan Cowen, a former researcher at Google DeepMind best known for his work in semantic space theory – a computational approach to understanding emotions and expression. The announcement came alongside news that the start-up had raised $50m in Series B funding.
ADD WORKGROUP