Google just launched MusicFX, an experimental tool that whips up custom music tracks from simple text prompts. Powered by MusicLM, the tool bakes DeepMind's SynthID watermarks directly into the audio to flag AI origin. To steer clear of legal trouble, it outright refuses to generate vocals or mimic specific artists. It is rolling out first to users in the US, Australia, New Zealand, and Kenya.
Meta just released an interactive demo for Audiobox, its new foundation research model for audio generation. The playground allows users to test zero-shot TTS, generate custom sound effects from text, and experiment with audio infilling. It gives creators an early hands-on look at Meta's generative audio capabilities.
Google DeepMind and YouTube just introduced Lyria, their most advanced music generation model yet. Alongside it comes Dream Track, a Shorts experiment that lets select creators generate 30-second tracks mimicking styles and vocals of artists like Charlie Puth and T-Pain, all watermarked via SynthID. It is a major push into licensed, generative audio with the music industry's blessing.
YouTube just announced new AI guidelines allowing record labels to take down tracks that mimic an artist's singing or rapping voice. Starting next year, creators must also label realistic AI-generated videos or risk demonetization. Music partners get a manual takedown form with zero parody exceptions, showing how far YouTube will bend to keep the music industry happy.
Meta just introduced Audiobox, an audio generation model that creates speech, sound effects, and ambient soundscapes using natural language prompts. It allows creators to combine voice samples with text prompts to restyle voices across different emotions and environments. The model is currently limited to select academic researchers alongside an interactive demo.
The Beatles just released "Now And Then," the legendary band's final original song featuring all four members. Using machine learning audio separation tech developed for Peter Jackson's Get Back documentary, the team cleanly isolated John Lennon's vocals from a murky 1970s cassette demo. It marks the first 21st-century Beatles track and shows how AI can revive lost musical history without generating synthetic voices.
The Beatles just announced that their final collaborative track, “Now and Then,” is dropping on November 2nd. Using the machine learning tech developed for Peter Jackson’s Get Back docuseries, producers successfully isolated John Lennon’s vocals from an unreleased 1970s demo cassette. It’s a remarkable showcase of AI unearthing lost music history without faking it.
Universal Music Group just sued Anthropic in Tennessee, alleging its Claude 2 model generates verbatim copyrighted lyrics like Katy Perry's "Roar" without a license. Music publishers argue Anthropic scraped and distributes protected tracks without paying for rights like traditional lyric sites, calling the lack of copyright guardrails outright infringement.
YouTube is developing an AI voice-cloning tool that lets creators mimic famous musicians while recording audio, according to Bloomberg. The streaming giant is currently in talks with major record labels to license their music catalogs for AI training. Why it matters: Following a wave of viral AI deepfakes, YouTube is trying to legitimize voice cloning into an officially licensed creator tool—if record labels agree to sign on.
Universal Music Group just teamed up with music creation app BandLab in a first-of-its-kind strategic AI alliance. The partnership focuses on ethical AI practices and protecting artist copyrights while exploring commercial opportunities for creators across both platforms. Why it matters: Major record labels are moving from fighting AI to actively setting ground rules with mainstream creation tools.
YouTube just announced new AI features for creators, including a smart music concierge and built-in automated dubbing. Creators will soon be able to type video descriptions into Creator Music to surface the right tracks within budget, while its Aloud integration inside YouTube Studio will generate multilingual voice tracks with one click.
Stability AI just launched Stable Audio, its first generative AI product for music and sound effects. Built on latent diffusion models trained on licensed AudioSparx tracks, it renders 44.1 kHz audio from descriptive text prompts and duration limits. A free tier generates up to 45 seconds of audio, while the Pro tier offers 90-second commercial tracks.
The Recording Academy just confirmed that the viral AI Drake song "Heart on My Sleeve" is not eligible for a Grammy. While human-written, CEO Harvey Mason Jr. clarified that its AI vocals were neither legally cleared nor commercially available. Why it matters: Mainstream music institutions are drawing a hard boundary against unlicensed voice cloning, no matter how viral the track becomes.
Ghostwriter, the anonymous creator behind the viral AI Drake track, just dropped a new song mimicking Travis Scott and 21 Savage while submitting his previous AI hit for Grammy consideration. Recording Academy leadership confirmed the track is creatively eligible because a human wrote it, though commercial distribution hurdles remain. The music industry can no longer dodge the AI voice controversy.
YouTube just launched a new Music AI Incubator alongside Universal Music Group, setting up three core principles for working with the music industry. The project ropes in artists like Ryan Tedder and Max Richter to guide generative AI experiments while scaling copyright safeguards like Content ID. Why it matters: YouTube is moving fast to get rights holders on board before AI music completely upends licensing models.
UC Berkeley researchers used brain recordings from 29 epilepsy patients to reconstruct a clip of Pink Floyd's "Another Brick in the Wall, Part 1" with 43% spectrogram similarity. By training AI to decode how the brain processes rhythm and pitch, the breakthrough could eventually make speech prosthetics sound genuinely expressive rather than robotic.
Google is reportedly negotiating with Universal Music Group and Warner Music to legally license artists' voices and melodies for a new AI song generator. The tool would let users make AI tracks while paying copyright holders, giving artists an opt-in choice instead of fighting unauthorized deepfakes.
Google just partnered with Grammy-winning lyricist Lupe Fiasco to launch TextFX, an AI tool built specifically for writers and rappers. Instead of writing verses for artists, it uses language models to generate alternate meanings, wordplay, and conceptual tangents from chosen words. The goal is to act as a brainstorming assistant rather than replace the artist.
Meta just open-sourced AudioCraft, a unified codebase designed to generate high-fidelity audio and music from simple text prompts. The release packages together MusicGen, the sound-effect generator AudioGen, and an improved EnCodec neural decoder, complete with open weights and code for researchers to train their own models.
Google just pulled its AI Test Kitchen app from both the App Store and Google Play, shifting the experimental playground entirely to the web. The platform, which once showcased early LaMDA demos, now only hosts the MusicLM text-to-audio tool. Google confirmed it is prioritizing the web experience to push updates faster on a single platform.
