Chen Gen： Generating music from words, Meta releasing new AI tools？

　　Text/Chen Gen

　　On Wednesday, US Eastern Time, Meta released an open source artificial intelligence (AI) tool, AudioCraft, through which users can create music and audio according to text prompts.

　　Meta said that this artificial intelligence tool combines three models or technologies, namely AudioGen, EnCodec and MusicGen, and can generate high-quality and realistic audio and music with text content. The simple understanding is that AIGC is generated according to the text content, but Meta focuses more on the intelligent voice direction.

　　This technology itself is not a new technology, and Baidu has already started to apply it earlier. However, from the current actual situation, there are still many problems, such as the obvious sense of machine when generating speech according to the text, and the lack of emotion and intonation expressed by people.

　　From the music direction, for Meta, it is the most consistent with the entertainment attributes of the platform, and it can also increase the user's stickiness. For entertainment and socialization, fun is the core.

　　With the help of literary instinct, music or audio can be generated for your own words, which will increase the entertainment interest, and also allow some people who may not be confident in singing to generate singing audio for themselves to share. At the same time, Meta can also take the opportunity to save a lot of music copyright fees.

　　However, the introduction of Meta technology will bring two challenges.

　　On the one hand, it is a challenge to the music industry, including music copyright. When AI has the ability to compose music and generate sound, it will have an impact on the pop music industry that takes fast food culture itself.

　　On the other hand, it will further affect the user's safety, especially when the sound can be generated according to the user's specific timbre. Some false fraud based on sound is a serious problem that cannot be ignored.