News JVTech Google invents a musical ChatGPT: here’s where to listen to this crazy AI
Published on 01/30/2023 at 15:50
To share :
In its (unofficial) war against ChatGPT, Google is looking for solutions to enter the artificial intelligence market. The latest, the presentation of a rather impressive musical AI.
Google launches chain projects to rival ChatGPT
2023 got off to a flying start for artificial intelligence! Between the ChatGPT phenomenon, the many image-generating AIs that are a hit on TikTok and the Tech giants that are starting to enter the dance, we are not about to see the AI madness abate. Speaking of tech giants, Google has also missed the train of artificial intelligence. To tell you, with the ChatGPT UFO, the GAFAM member even had to initiate a “red code” internally so as not to let the OpenAI tool take over. In fact, this chatbot is so advanced, it almost becomes a threat to Google and other web tools like Wikipedia..
By this “red code”, it should be understood that Google – or rather its parent company Alphabet – mobilized several engineers to develop a tool comparable to that of OpenAI. And it was Demis Hassabis, the CEO of Deepmind (a subsidiary of Alphabet) who announced that the Sparrow project was under development and that its launch should take place in 2023. But that’s not all, since Google also shared, a few days ago, a communiqué on a new artificial intelligence called MusicLM.
Music: Google’s new AI that generates music from texts
As indicated by his name, MusicLM is a model capable of generating music tracks from textual descriptions. But that’s not all, Google’s AI can also transform a whistled or hummed melody into a real musical composition. In their communiqué, the teams in charge of the project offer several excerpts. We can thus listen to a melody created from two words “melodic techno”, another created from a whistle or even one created from the description of a place/situation “beach in the caribbeans”.
Screenshot: Google Research
In addition to all these features, MusicLM is also able to simulate a human voice, whether in tone or sound. If you’ve listened to some snippets with vocals, I think you’ll agree with us that it’s not perfect yet. However, and unlike other AIs of the genre (Jukebox or AudioML), MusicLM is the only one to have such a large database. And if Google continues to enhance it, the voices should gradually become much more qualitative.
If ChatGPT is winning the bet on the side of chatbots, MusicLM could be the next big AI to be talked about. Google has not yet indicated a release date or official launch for its new model, but if the Tech giant wants to compete with OpenIA and the other GAFAMs, we could have news of this project very soon.
