Mistral, French start-up of artificial intelligence (IA), launched its first models on Tuesday focused on vocal recognition and transcription in several languages. “The voice will be crucial in the future of human-machine interactions and will play a critical role in the adoption of artificial intelligence,” the company told AFP.
This open source model, called VOXTRAL, makes it possible to transcribe audio content, live or from imported files, in several languages ranging from English to Hindi, automatically recognized. It can also make summaries, responding to requests posed orally and Mistral intends to add other features soon such as the recognition of several interlocutors and their characteristics (age, sex) but also their emotions, according to a press release. Voxtral can in particular be used to improve business vocal systems to respond to their customers by phone, according to the start-up. The French company also develops with the automaker Stellantis a system allowing drivers to interact orally with an AI assistant on their vehicle.
The American mastodon Openai presented a vocal mode for its GPT-4O model, capable of “reasoning” in real time via audio, vision and text. This version of Chatgpt can notably read users’ emotions on faces via the camera of a smartphone.
The French Research Laboratory in Artificial Intelligence Kyutai, founded by Xavier Niel, owner of the Iliad group, and Rodolphe Saadé, CEO of the Maritime Transporter CMA CGM, unveiled in February a simultaneous translation model. Called “hibiki” (“echo” in Japanese), this AI translates the words of a real -time user from French to English, as an interpreter would do.
Mistral, French start-up of artificial intelligence (IA), launched its first models on Tuesday focused on vocal recognition and transcription in several languages. “The voice will be crucial in the future of human-machine interactions and will play a critical role in the adoption of artificial intelligence,” said the company. This model in open source, baptized by Voxtral, thus makes it possible to transcribe audio content, live or from imported files, in several languages ranging from English to Hindi, automatically recognized. He can also make summaries, responding to requests for oral and Mistral intends to add other features soon such as the recognition of several interlocutors and their characteristics …
