ChatGPT gets a new competitor: Say hello to Moshi, which can understand your tone of voice

Just a few days ago, OpenAI made headlines for delaying ChatGPT’s long-awaited voice mode. The company had stated that it had to fix technical issues and ensure a high quality standard. This was somewhat of a disappointment for everyone who was looking forward to talking to the AI ​​chatbot. But what if we told you that there is another chatbot that can not only talk to you but also understand your tone of voice? Say hello to Moshi, developed by a French AI company called Kyutai.

Read more

Say hello to Moshi

Moshi is an AI voice assistant that enables lifelike conversations similar to Amazon’s Alexa or Google Assistant, leveraging the powerful Helium 7B voice model. The new chatbot stands out for its ability to speak in different accents and use 70 different emotional and speaking styles. It can also understand the tone of your voice as you speak to it. In addition, Moshi can process two audio streams simultaneously, allowing it to listen and respond at the same time. The launch of the voice assistant was recently live-streamed and has been making headlines ever since.

According to a report in Tech Radar, developing Moshi required an extensive fine-tuning process that involved over 100,000 synthetic dialogues created using text-to-speech (TTS) technology. To improve the chatbot’s voice quality, Kyutai worked with a professional voice actor to ensure Moshi’s responses sounded natural and engaging.

“This new type of technology makes it possible for the first time to communicate with an AI in a smooth, natural and expressive way,” the company said in a statement, according to Tom’s Guide.

Available now

You can try Moshi for yourself right now as a demo version is available. Just go to us.moshi.chat and follow the instructions. At the moment, you can talk to the AI ​​voice assistant for a maximum of 5 minutes.

Before you interact with Moshi, you’ll come across a message that reads: “Moshi is an experimental conversational AI. Take everything she says with a grain of salt. Conversations are limited to 5 minutes. Moshi thinks and speaks simultaneously. Moshi can listen and speak at any time: maximum flow between you and Moshi. Ask her to play a pirate RPG, make lasagna, or say what movie she last saw. We strive to support all browsers, Chrome works best. Baked with <3 @Kyutai You are on the US demo. Depending on your location, the EU demo may offer better latency."

What’s next?

Kyutai aims to make Moshi an open-source project. By sharing the model’s code and framework, the company hopes to encourage innovation and address ethical concerns surrounding the development of artificial intelligence. This open-source strategy has the backing of prominent supporters, including French billionaire Xavier Niel.

In the future, Kyutai plans to integrate advanced features into Moshi, such as AI audio identification, watermarking, and signature tracking systems. These additions will help ensure accountability and traceability for AI-generated audio, promoting transparency in AI technology.

If Moshi gains traction, it could serve as a catalyst for other voice-powered AI assistants and accelerate the adoption of large language models into existing systems like Alexa. The impressive capabilities demonstrated by Moshi suggest a promising future for voice AI technology.