The American start-up Hume is launching a voice interface capable of empathy, powered by a very specific language model. This form of conversational artificial intelligence generates voice responses adapted to the mood of the human with whom it interacts, to offer a much more natural experience than any other voice assistant.
Hume introduces Empathetic Voice Interface (EVI), a first-of-its-kind conversational AI with emotional intelligence that can analyze voice tones to understand when users finish speaking and optimize responses to achieve user satisfaction.
Rather than having mechanical and unnatural responses, a bit like when you use the ChatGPT oral tool, here you are entitled to a truly immersive conversation, which takes into account both the content and the form when you address to her. Every time you speak to it, the interface analyzes your voice, in terms of determination, interest, concentration, boredom, calm or satisfaction and responds with the appropriate tone. The result gives the impression of actually chatting with someone. The most curious can already try it (in English only): demo.hume.ai.
If it is already available for testing, this technology will be truly launched in April 2024. Developers will then be able to integrate it into their projects. Its uses are relatively simple to understand. This type of empathetic voice interface could, for example, be used in customer services and various call centers. Imagine an after-sales service capable of responding to you with attention and empathy. More specifically, it could also be useful in the field of mental health, for dialogue with more or less depressed patients.