What the announcement includes and leaves out

Google is rolling out Lyria 3.5, its music generation model, inside Flow Music. The company says it builds richer, more complex melodic structures that sound more natural, improves prompt adherence and structural awareness in lyrics, delivers more realistic and emotionally nuanced vocals with better pronunciation, and gives more control over tempo and duration.[1]

The same page carries no comparative evaluation result, no country or language list, no subscription tier, no statement of whether this is a preview or general availability, and nothing about whether generated tracks are watermarked. What Lyria 3.5 comes with is a statement about availability; it does not tell someone opening their account today whether they will see the feature.[1]

The destination is getting narrower

On the same day Sony Music, Universal Music Group and Warner Music Group, with smaller labels and independents, asked chart companies to include only primarily human-made tracks, to require AI involvement to be demonstrably lawful, and to keep music that could be committing streaming fraud out of the charts. The report names Billboard and Spotify's top-song lists, says tens of thousands of AI-generated songs are uploaded daily and notes Spotify deleted more than 75 million tracks last year.[2]

The order matters: a tool like Lyria 3.5 improves while the rule deciding where its output counts is still unwritten. No chart operator's answer appears in the report, so it is not possible today to know whether a track made in Flow Music would be excluded, labelled or untouched. Another reading is available: the request from Sony Music and the others may in practice narrow to tighter fraud screening, in which case an individually made track is unaffected.[1], [2]

Three signals worth waiting for

First, a published price and a country-and-language list for Flow Music. Second, a plain sentence on whether output carries a mark showing where it came from. Third, a written policy from Billboard or Spotify. Until those three arrive, more realistic vocals stays a product feature and is not enough to tell where a track is allowed to go.[1], [2]