AI Music

Factual summary

Gerencia 360 Music filed a copyright complaint alleging that Suno copied its regional Mexican catalogue from streaming services and used it to train models capable of generating Spanish-language songs. The company is reportedly seeking nearly $100 million. The allegations have not been adjudicated.

Factual summary

Anthropic gave a substantive response to the Sony Music Publishing and Warner Chappell complaint, characterizing it as recycled litigation, promising a robust defense and expressly maintaining that AI training makes fair use of copyrighted material. The underlying complaint remains unresolved.

Factual summary

Sony Music Publishing, Warner Chappell Music and affiliated publishers filed a federal complaint against Anthropic and co-founders Dario Amodei and Benjamin Mann. The plaintiffs allege large-scale torrenting, scraping and copying of copyrighted musical works for Claude, removal of copyright-management information and reproduction of protected lyrics. The allegations have not been adjudicated.

Factual summary

Approximately 80 performers backed the Save Our Voices Now campaign calling for UK legislation giving individuals statutory control over their voices and protection against unauthorized AI voice cloning. The campaign includes an open letter and public petition and does not create new law.

Factual summary

KOMCA voted to withdraw amendments governing registration and trust treatment of AI-assisted music. New registrations will be withheld while broader standards are discussed, and Yonhap reported that works already accepted under the withdrawn framework will enter cancellation procedures.

Factual summary

Stability AI raised $76 million in Series B capital with Universal Music Group, Sony Music Group and Warner Music Group among the investors. Stability says the funding will support creative-production products, applied research and professional services.

Factual summary

Luminate built a framework using streaming data, metadata and audio-identification technology to flag AI-generated songs and artists. It says thousands of items have been identified, while user-visible labels and source-attribution integrations remain future rollout steps.

Factual summary

Soundplate acquired Stemmer and brought its mastering algorithms and B2B and B2C operations in-house. It rebuilt the technology around a proprietary serverless mastering API serving its mastering, vinyl and stem-processing tools.

Factual summary

Music IP Holdings opened a generative-AI patent licensing framework and identified Udio and GRAI as its first technology licensees. MIH says its portfolio includes more than 24 issued or allowed patents and more than 50 pending applications spanning prompt handling, moderation, watermarking, identifier tagging, authorization, licensed distribution and payment.

Factual summary

Apple told industry partners that its existing AI Transparency Tags will become consumer-visible Made With AI labels later in 2026. Labels and distributors will be required to tag content when AI generated a material portion of a track, composition, artwork or music video.

Factual summary

Adobe moved Firefly Generate Music into general availability. The feature creates custom-duration instrumental music using user controls or video-derived guidance, and Adobe describes the model as trained on licensed audio and its outputs as royalty-free, fully licensed and commercially safe.

Factual summary

Round Hill Music and affiliates filed separate Northern District of California copyright cases against Suno and Anthropic concerning alleged use of protected songs in AI training. Each complaint begins with 500 representative compositions and says the asserted scope may expand to 10,000 or more works.

Factual summary

Alibaba launched HappyShrimp 1.0 in beta for desktop web users in China and internationally. The full-song model generates lyrics, melody, arrangement and vocals from natural-language instructions, while Taihe Music Group will participate in artist co-creation, content development and AI music platform cooperation.

Factual summary

Spotify and Kobalt announced a licensing agreement for Spotify's forthcoming fan-made covers and remixes product. Participating Kobalt songwriters will be able to share in value generated through licensed derivative creations, expanding the publishing-rights coverage being assembled around the paid Premium add-on.

Factual summary

Suno launched Studio 2.0 for Premier subscribers with MIDI import, recording and editing; MIDI-to-audio prompting; a session-aware beta chat agent; advanced stem separation; custom plugin generation; automation; and unlimited 32-bit/48kHz multitrack and stem exports from Studio.

Factual summary

Beatport expanded its Beatdapp partnership to identify AI-generated music. Fully or majority AI-generated tracks will be withheld during ingestion, while majority-human AI-assisted releases remain eligible and are tagged for Beatport's curation team.

Factual summary

Thailand's Department of Intellectual Property said protected songs used for AI voice conversion, remixes or reference-based generation may implicate separately controlled reproduction, adaptation and communication rights. It advised users to identify and secure the relevant permissions before use.

Factual summary

BMG and Suno announced a global strategic alliance establishing an opt-in framework covering BMG's recorded-music and music-publishing repertoire. BMG says participating artists and songwriters will have their rights protected and be compensated, while the agreement also settles prior use of covered recordings and publishing works. BMG will participate in development around Suno's forthcoming music-industry-partnered model.

Factual summary

Spotify has opened AI Persona self-disclosure and will begin displaying badges in mid-September for artist profiles whose public identity may be AI-generated. Spotify may also designate profiles after review, artists can appeal, and AI Personas will be excluded by default from editorial and algorithmic recommendations.

Factual summary

D'Addario said an internal review confirmed that Suno Studio was used to regenerate music in its NYXL HD promotional demo, reversing earlier explanations. The company apologized for sharing inaccurate information and said employees and creative partners will be required to disclose generative-AI use.

Factual summary

Suno has specified monthly song-download limits beginning September 3, 2026: 20 downloads for Pro subscribers and 60 for Premier subscribers, while the Free plan lists no recurring monthly song downloads. Paid plans continue to advertise commercial-use rights for new songs made while subscribed.

Factual summary

Suno announced a package of transparency and platform-integrity changes for generated music. The company says it is starting to roll out off-platform transparency tools, plans to adopt audio watermarking and fingerprinting technology, will introduce a download policy intended to limit mass distribution to streaming services, and has made its rules against unauthorized song recreation, rightsless uploads, voice or likeness misuse, fake engagement, bots and deceptive audio more explicit.

Factual summary

Treblo, formerly Sonauto, published an open-source classifier that estimates whether an audio file was created with Treblo. The local command-line tool returns a Treblo probability, label and confidence level and uses a downloadable model checkpoint. Treblo also launched a browser-based detector.

Factual summary

Reuters reported that Spotify announced a new agreement with Merlin for Spotify's upcoming paid tool for fan-made covers and remixes. The agreement will allow artists on labels participating in Merlin's Spotify agreement to take part in the product.

Factual summary

The Munich Regional Court I issued a first-instance judgment largely granting GEMA's claims for injunctive relief, information and damages concerning six musical works. The court found actionable reproduction connected with United States training, model memorisation in Germany and generated outputs. The judgment is not final.

Factual summary

The European Commission states that Article 50 transparency obligations for relevant AI providers and deployers apply from August 2, 2026. Providers of systems generating synthetic audio and other content must use machine-readable marking, while deployers must disclose defined deepfake audio, image or video content.

Factual summary

IFPI announced that it is applying AI chart-eligibility principles across the official charts it directly manages in Latin America, the Middle East, Africa and Southeast Asia, including the Official Philippines Chart, while working with its national-group network to expand the programme across more than 20 additional chart programs.

Factual summary

Google launched Lyria 3.5 in Flow Music with reported improvements to musicality, lyrics, prompt adherence, vocal expression, pronunciation, tempo control and output duration. Google DeepMind describes generation of tracks up to three minutes and states that generated or edited outputs carry an imperceptible SynthID watermark.

Factual summary

Eleven major and independent music companies jointly proposed global principles under which recordings developed with generative-AI services would qualify for official charts only when the AI services are authorised and lawful, the recordings are substantially human-made, manipulation concerns are absent, applicable rights laws and service terms are respected, and AI involvement is disclosed downstream.

Factual summary

HYBE publicly confirmed that it will liquidate AI audio subsidiary Supertone. Supertone's shareholders resolved to dissolve the company and appoint a liquidator on July 15. In its July 27 public response, HYBE said it had reviewed alternatives, concluded that continued operation would be difficult, and would continue pursuing AI capabilities through collaboration with specialist technology partners.

Factual summary

GEMA launched PLAI by GEMA, a configurable dataset for training AI music tools that bundles audio files, detailed metadata, composition rights, and master rights. GEMA states that the current package contains approximately 57,000 works and 178,000 audio files across more than 60 genres. Klangio is the first customer, and contributing rightsholders are intended to receive proportional and appropriate compensation.

Factual summary

Concord Music Group, Universal-affiliated music publishers, and ABKCO Music filed a 71-page second amended complaint against Anthropic in the Northern District of California. The publishers allege that Anthropic copied protected lyrics during model development and operation, generated identical or near-identical lyrics in Claude outputs, and removed or omitted copyright-management information. The complaint is a party allegation and does not establish liability.

Factual summary

SOCAN and Musical AI announced a collaboration to explore opt-in consent, attribution, reporting, licensing, and compensation infrastructure for generative-AI music. SOCAN will recognize Musical AI as an approved technology partner for attribution services and related initiatives. The announcement establishes a collaboration and intended direction but does not state that a production system has already been deployed.

Factual summary

Sony Music Entertainment and nine affiliated record companies filed a second federal complaint against Udio asserting 30,117 sound recordings. The plaintiffs allege that Udio copied the recordings without permission for generative-AI model training and obtained many recordings by stream-ripping them from YouTube. The complaint contains allegations and requests for relief; it does not establish liability.

Factual summary

Suno filed a response opposing Universal Music Group's and Sony Music Entertainment's effort to obtain its settlement and licensing agreement with Warner Music Group. Suno argued that litigation-driven licenses have limited value as evidence of a functioning market price for AI training and that disclosure could compromise settlement dynamics and discourage future settlements.

Factual summary

404 Media reported that hacked historical Suno code from 2023 and 2024 appears to document scraping instructions, acquisition sources, dataset categories, and quantities involving YouTube Music, Deezer, Genius, and other platforms. Suno acknowledged a November 2025 security incident, characterized the exposed code as outdated, and reiterated that its models were trained on publicly accessible music files and metadata.

Factual summary

SoundPatrol announced on July 16 that its Unified Detection and Review System had passed the C2PA Conformance Program as a validator. The official C2PA Conformance Explorer independently lists the product and identifies January 28, 2026 as its latest conformance date, so the July event is treated as an announcement rather than the date conformance was first achieved.

Factual summary

Luminate's 2026 Midyear Report states that 54 percent of surveyed U.S. musicians expressed positive feelings or acceptance toward generative-AI music tools, compared with 35 percent of non-musicians, and that 18 percent of U.S. musicians use AI to edit or remix music, compared with 6 percent of non-musicians. Its highest-ranking AI-assisted song was number 282 globally with 210.7 million global audio streams during weeks 1 through 24 of 2026.

Factual summary

Music Business Worldwide reported that Suno and the Digital Media Association responded to a voluntary industry proposal for AI-Generated and AI-Assisted track-level labels for sound recordings. Suno supported transparency while emphasizing work with artists, rightsholders and platforms, and DiMA emphasized reliable metadata across the music supply chain. The report did not establish adoption of the proposed labels by a major digital service.

Factual summary

The Guardian reported that producers and an RMIT researcher questioned whether Josh Fawaz's cover of Like a Prayer involved generative AI. Fawaz said he uses AI as a tool, but the reporting and published credits did not establish the material extent of that use. The track reached number one on Australia's National Radio Airplay chart, while the commercial-radio disclosure requirement for AI-generated voices in programming does not apply to music.

Factual summary

A fictional rights platform published a sample framework describing granular creator permissions for model training, synthetic voice use, and commercial outputs.

Factual summary

A fictional streaming service proposed new metadata fields for identifying materially generated audio and authorized synthetic performances.