The first half of October 2026 has been one of the busiest stretches yet for AI in music. In less than two weeks, a streaming service handed listeners an off switch for AI-generated tracks, a major label put a new number on voice deepfakes, a US court handed down a prison sentence for AI-powered streaming fraud, and new production tools brought AI further into the everyday studio.
None of these stories is about whether AI belongs in music. That debate is over in practice. They are about who gets to decide how AI music is labelled, paid and used. Here is what happened and why we think it matters.
Streaming platforms give listeners the choice
On October 13, Deezer announced an on/off toggle that lets users exclude fully AI-generated tracks, albums and related artists from search results. The setting will roll out globally before the end of 2026, and it can be switched off at any time.
The numbers behind the decision are striking. According to Deezer:
it receives nearly 90,000 fully AI-generated tracks every day, more than 50% of total daily deliveries on peak days;
it detected and removed more than 42 million AI-generated tracks from algorithmic and editorial recommendations between January 2025 and August 2026;
fully AI-generated music accounts for only 1-3% of streams, but up to 85% of those streams in 2025 were found to be fraudulent.
Deezer also stresses that the filter targets fully AI-generated music. Tracks with limited AI-assisted elements inside a broader human creative process are treated differently. TechCrunch and Music Business Worldwide also covered the move.
That distinction matters for producers. Using AI as a tool and uploading machine-made tracks at industrial scale are now treated as two separate things.
Deepfakes at industrial scale
On October 5, Music Business Worldwide reported, citing the Financial Times, that by the end of September Sony Music had asked digital platforms to remove more than 260,000 AI-generated tracks impersonating its artists. That is almost double the 135,000 it had flagged by the end of March. Artists targeted include Harry Styles, Adele, Britney Spears, Queen and Michael Jackson.
Sony told the FT it expects platforms "to act quickly and to prevent the same content from reappearing," adding: "it's becoming an uphill struggle." For working artists, the lesson is simple. Your voice and likeness are now assets that need active protection, not just your recordings.
Streaming fraud meets the courts and regulators
A first prison sentence
On October 6, the US Attorney's Office for the Southern District of New York announced that Michael Smith had been sentenced to 18 months in prison for a streaming fraud scheme that ran from 2017 to 2024. Prosecutors say he created hundreds of thousands of AI-generated songs, streamed them billions of times through as many as 10,000 bot accounts at once, and collected more than $8 million in royalties. He was also ordered to forfeit $8,091,843.64. Billboard noted that the sentence caps off the first-ever criminal case over streaming fraud.
Washington asks for evidence
A day later, on October 7, the US Copyright Office published a notice of inquiry on music streaming fraud, following a congressional request. Among its questions, the Federal Register notice asks what data exists on the link between streaming fraud and AI-generated music, and how it affects royalty licensing and distribution. Initial comments are due November 23, 2026, and reply comments by December 21, 2026.
Every fraudulent stream dilutes a shared royalty pool. Stronger enforcement is, in the end, a way of protecting real listeners and real creators.
Licensed models, fresh scrutiny
The licensing era is also moving forward. Speaking at Bloomberg Screentime, Suno CEO Mikey Shulman said the company is now "far beyond" two million subscribers and $300 million in annual revenue, Digital Music News reported. This follows Suno's September switch to v6 models built in partnership with Warner Music, BMG and Believe.
Licensing does not end the questions, though. On October 6, Italy's competition authority (AGCM) opened an investigation into Suno's terms of service. It flagged unilateral changes to prices and service, account terminations, arbitration and a broadly framed content license that requires users to waive their moral rights. The authority said the copyright provisions "deserve particular attention". A public consultation is expected in the coming weeks.
For anyone making music on an AI platform, the terms of service are now part of your creative contract. Read them as carefully as you would a label deal.
Tools that fit into the studio
Away from the courtrooms, AI keeps getting more practical for producers:
Google Flow Music now lets producers export custom instruments and effects built with natural language as VST3/AU plugins, so they run inside a standard DAW (October 6).
Suno Studio, the company's browser-based DAW, added better downbeat detection, clip time-stretching, curved fades and pop-out windows (October 8).
SubverseLab ELEK is a new stem separation app and plugin that splits vocals, drums, bass and other parts using HT-Demucs, runs entirely on your computer with nothing uploaded, and also reads key and BPM (October 12).
The common thread: AI is showing up less as a "make me a song" button and more as a set of tools that sit inside the producer's existing workflow.
What it means for creators
At Xdripia Waves we see October as a turning point from hype to accountability. The tools are getting better and more personal, while platforms, labels and regulators are building the rules around them at the same time. For creators, a few practical takeaways stand out:
Be transparent. Labelling and detection are becoming standard. Clear disclosure of how AI was used protects your reach and your reputation.
Protect your identity. With deepfakes counted in the hundreds of thousands, keep an eye on your artist profiles and on releases delivered under your name.
Read the terms. Who owns your outputs and which rights you give up now depends on the platform.
Use AI as an instrument. The most useful tools, from local stem separation to custom plugins, extend human craft rather than replace it.
We will keep following these stories and testing the tools that reach the studio. In our view, the most valuable thing in music is still a human idea worth hearing. AI is changing how we get there, not why.