Alphabet Inc. Class C(GOOGL.US) ignites the AI battle again: Gemini launches music model, generating songs with just one sentence.

On Wednesday (February 18), in the midday trading of the US stock market, Google announced on its official blog that the Gemini app is now equipped with the company's most advanced music generation model, Lyria 3, allowing users to create 30-second music clips using text or images. The press release stated, "Just describe an idea or upload a photo, and Gemini can turn it into a high-quality, catchy song in seconds. To further expand creative space, you can even let Gemini draw inspiration from the uploaded content." According to Google, Gemini can not only "turn text into a song," but also create a song that perfectly fits the atmosphere by "uploading a photo or video," with the song being 30 seconds long and having a custom cover designed by Nano Banana. Additionally, creators can explore Lyria 3 through YouTube's Dream Track feature to enhance the quality of music for Shorts short videos. Google stated that Lyria 3 has been launched in the United States and is gradually being introduced to creators in other regions. Google mentioned that Lyria's audio generation function incorporates the company's latest privacy and security features, with all generated tracks containing imperceptible watermarks added using SynthID technology to detect if the music was created or edited by artificial intelligence. After the announcement, the stock price of the mainstream European and American streaming music service platform Spotify (stock code: SPOT) rolled back nearly 5% of its intraday gains, and Sirius XM (stock code: SIRI) also briefly turned negative. Some analysts believe that while Google's music generation model is not expected to be a fatal blow to Spotify, these actions may force the latter to quickly launch AI mixing functions. In addition, adding audio creation tools to mobile apps is expected to enhance Google's consumer product competitiveness. For Google, the company needs to continue proving to investors that its AI product investments can lead to revenue growth. Generative AI tools have always been treated cautiously or even hostilely in the music industry, with many practitioners believing they could threaten business models and intellectual property. Google emphasized in the blog that its system has protective measures in place to prevent AI from directly appropriating specific artists' works. If users nominate real musicians, Gemini will only consider them as "broad sources of creative inspiration," generating tracks with styles or atmospheres similar to theirs. This article is reprinted from Cailianshe, GMTEight editor: Chen Wenfang.