Updated Oct 29

AI Copyright: Who Owns the Training Data

The fair-use question for chatbots is fully argued in New York and still unanswered anywhere. On 23 October the publishers, authors, OpenAI and Microsoft filed their oppositions in the consolidated cases; replies are due 20 November. In San Jose, Judge Eumi Lee has not ruled on the music publishers’ lyrics case she heard on 21 October.

The New York briefs put new facts on the record about where the books came from. OpenAI says only GPT-3 and GPT-3.5 trained on books downloaded from LibGen, that it stopped in late 2021, and that Books3 trained none of the models at issue. Microsoft says it downloaded nothing. The publishers argue that copyright is not an opt-out system.

Music owners keep filing. Reservoir sued Anthropic over 853 songs’ lyrics and Suno over 1,050 compositions on 21 October, seeking about $338 million. Two days later Elvis Presley’s song owners sued Anthropic over 274 songs, seeking about $41 million. Each figure is the statutory maximum, and none has been tested.

London’s Court of Appeal hears Getty v Stability AI from 10 November; Delhi hears ANI’s appeal on 8 December.

New Oct 29

The whole story5 chapters

The state of it — as of October 29, 2026

Item	Status	Figure	As of
Microsoft Copilot Copyright CommitmentVendor indemnity: Microsoft absorbs commercial customers' copyright exposure over Copilot output	Offered	$0 to customer	Sep 7, 2023
Bartz v. Anthropic (pirated books)Authors' class action over pirated books used to train Claude; No. 24-cv-5417	Settled	$3,000 / work	Jul 21, 2026
Bartz v. Anthropic (plaintiffs' fees)Court-set price of winning the pirated-books case	Awarded	$101.56M fees	Jul 21, 2026
Concord, UMPG & ABKCO v. Anthropic (lyrics)Music publishers' statutory-damages claims over song lyrics used to train Claude	Sought	$150,000 / work max	Jul 22, 2026
GEMA v. SunoDamages for training on and reproducing six protected works; 42 O 763/25	Amount pending	Not yet set	Jul 31, 2026
Round Hill Music v. Suno and v. AnthropicStatutory damages over a music catalogue used to train two model builders; reported Nos. 5:26-cv-08507 (Suno) and 5:26-cv-08505 (Anthropic)	Sought	Up to $1B / case	Aug 17, 2026
BMG — Suno global allianceOpt-in licence over BMG's recorded and publishing repertoire, settling Suno's prior unlicensed use	Settled	Terms undisclosed	Aug 12, 2026
USA Today Co. v. OpenAINewspaper group’s damages demand over articles from 19 publications used to train and run ChatGPT; reported No. 1:26-cv-08892	Sought	Over $250M	Oct 8, 2026
Individual authors v. Meta (books)Five near-identical non-class suits by about 292 authors over books used to train Llama and Muse; lead case Winspear v. Meta, No. 3:26-cv-12100	Sought	Up to $150,000 a work	Oct 19, 2026
Reservoir v. Anthropic (lyrics)853 compositions whose lyrics allegedly trained Claude; No. 5:26-cv-12308	Sought	About $151M	Oct 21, 2026
Reservoir v. Suno (recordings and compositions)1,050 compositions allegedly scraped from streaming platforms to train Suno; No. 3:26-cv-12305	Sought	About $186M	Oct 21, 2026
Presley song owners v. Anthropic274 compositions recorded by Elvis Presley; ABG EPE IP LLC et al. v. Anthropic, No. 5:26-cv-12473	Sought	About $41M	Oct 23, 2026

Oct 23, 2026 · latest

What a work is worth: settlement figures, statutory ceilings, and fee awards, matter by matter.

Earlier this month

NewOct 23, 2026

Elvis Presley’s Song Owners Sue Anthropic

Authentic Brands Group’s Presley company, the catalogue’s publishers and family trusts sued Anthropic in San Jose on 23 October (No. 5:26-cv-12473) over 274 compositions Presley recorded. They seek up to $150,000 a song, about $41 million. The complaint alleges the lyrics reached Claude through lyric sites, Common Crawl and The Pile. It quotes Anthropic’s own admission in the Bartz case that a released Claude model trained on YouTube subtitles from a version of The Pile. Its evidence that Claude reproduces lyrics is drawn from earlier suits, not shown for these songs. These are allegations; Anthropic says it will defend.

Read

ChatGPT Is Eating the WorldThe GuardianBBC News

NewOct 23, 2026

OpenAI Tells the Court Only Its Older Models Trained on Pirated Books

Fair-use opposition briefs were filed on 23 October in the consolidated cases against OpenAI and Microsoft in New York (No. 1:25-md-03143). OpenAI said only GPT-3 and GPT-3.5 among the models at issue trained on books downloaded from LibGen. It says it stopped using them in late 2021 and never trained those models on Books3. Microsoft said it downloaded nothing and could not supervise OpenAI’s data gathering. The Times and other publishers argued that copyright “is not an opt-out regime”. They cited Microsoft data showing 83–93% click-through falls for their sites. These are the parties’ accounts; replies are due 20 November.

Read

CourtListener / RECAP document archive (Free Law Project)CourtListener / RECAP document archive (Free Law Project)CourtListener / RECAP document archive (Free Law Project)CourtListener / RECAP document archive (Free Law Project)Daily Cortex

NewOct 21, 2026

Reservoir Sues Suno and Anthropic Over About 1,900 Songs

Reservoir Media and Shapiro, Bernstein sued Anthropic (No. 5:26-cv-12308, 853 compositions) and Suno (No. 3:26-cv-12305, 1,050 compositions) in San Francisco on 21 October. Both complaints plead infringement, circumvention of access controls and removal of copyright notices. Statutory damages at the $150,000 ceiling plus the DMCA awards come to about $151 million against Anthropic and $186 million against Suno, roughly $338 million in all. The Suno complaint relies on leaked Suno code, reported by 404 Media, that allegedly shows scraping of YouTube Music, Deezer and Genius through Bright Data. These are allegations; Anthropic will defend, and Suno declined comment.

Read

ChatGPT Is Eating the WorldChatGPT Is Eating the WorldMusic Business WorldwideDigital Music NewsBillboardLaw Commentary

NewOct 19, 2026

Hundreds of Authors Sue Meta One by One, Rejecting Class Settlements

About 292 authors filed five near-identical suits against Meta in San Francisco federal court on 19 October (lead case Winspear, No. 3:26-cv-12100). Each sues individually so a jury sets damages per work: up to $150,000, against the roughly $3,000 a book paid in Anthropic’s settlement. They allege Meta torrented books from LibGen, Z-Library and Anna’s Archive, and that its later buying and scanning of print books for its Muse models is itself infringement, which Judge Alsup held fair use for Anthropic. On 26 October one was referred to Judge Chhabria to decide whether it belongs with Kadrey. These are allegations.

Read

ChatGPT Is Eating the WorldBloomberg LawPublishers WeeklyChatGPT Is Eating the WorldJustia (court documents)

Oct 16, 2026

A Less-Redacted Brief Says Microsoft Built a Secret Crawler for OpenAI

A partly unsealed summary-judgment motion from the news publishers in the consolidated OpenAI litigation (S.D.N.Y., 25-md-3143) says OpenAI paid Microsoft to build and run a web crawler, code-named “Project Mango”. Its copies, the motion says, later went into training models including GPT-4o, and Microsoft told employees to keep the project “secret”. The publishers say it took in more than 160,000 of their works; the payment and the number of pages remain sealed. Microsoft had withdrawn some redaction requests on 9 October. This is the plaintiffs’ account of sealed evidence as reported by Bloomberg Law; the filing itself was not retrievable.

Read

Bloomberg Law

Oct 15, 2026

Korean Books Turn Up in Anthropic’s Settlement, and Its Records Show a Hunt for Asian Print Books

At least 22 books by Korean authors, three of them by the Nobel laureate Han Kang, are among the 482,460 works eligible for payment in the Bartz v. Anthropic settlement, two Seoul newspapers found by searching its claims website; a work qualified only if it had been registered with the US Copyright Office. Internal Anthropic meeting records filed in the case show that by November 2024 it had agreed terms to buy print books from Korea’s largest wholesaler, Woongjin Booxen, had a supply deal with Japan’s Nippan, and was seeking Taiwanese and Chinese books. Woongjin confirmed the sales; JoongAng estimates 250,000–300,000 Korean books.

Read

Korea JoongAng DailyThe Chosun Ilbo

Oct 8, 2026

USA Today Sues OpenAI for More Than $250 Million

USA Today Co. and newspapers it owns sued OpenAI in the Southern District of New York (reported as No. 1:26-cv-08892), alleging hundreds of thousands of articles from 19 publications were copied to train and run its models. The complaint pleads infringement, vicarious infringement and removal of copyright notices, and seeks more than $250 million — reportedly up to $150,000 a work — plus an injunction. Microsoft is not reported as a defendant. One docket tracker counts 29 US copyright suits against OpenAI, against 17 for Meta. These are the complaint’s allegations as reported; OpenAI had not responded.

Read

The Trademark Lawyer MagazineHoodlineChatGPT Is Eating the World

Oct 6, 2026

Canada’s Largest News Publishers Sue Anthropic in Ontario

Postmedia, Torstar, The Globe and Mail and The Canadian Press sued Anthropic in an Ontario court on 6 October, reported on 19 October. The statement of claim alleges that Anthropic scraped their websites to train Claude and keeps scraping them to answer users’ prompts, breached the sites’ terms of use, circumvented technical measures meant to stop copying, and was unjustly enriched. No damages figure, court file number or response from Anthropic has been reported. The same publishers, with CBC/Radio-Canada, sued OpenAI in Ontario in November 2024; that case is unresolved. These are allegations.

Read

CBC NewsThe Globe and MailToronto StarCBC News

Oct 2, 2026

We intend to seek review by the Supreme Court to obtain much-needed clarity on these issues, which carry significant implications for innovation and the development of AI technologies.

— ROSS Intelligence

Not yet checked

No petition has been filed. The statement, made through ROSS’s counsel three days after it lost in the Third Circuit, names no date; it is read here as a petition docketed by 31 March 2027. Asking is not being heard — the Court chooses its cases.

Read

ChatGPT Is Eating the World

Oct 2, 2026

Courts Let YouTube Creators Sue Over How AI Companies Downloaded Their Videos

Judge Jacqueline Scott Corley in San Francisco refused to dismiss Ted Entertainment v. ByteDance (No. 3:25-cv-10933), a suit that skips fair use entirely. Its one claim, under the DMCA’s anti-circumvention rule (§1201(a)), is that ByteDance bypassed YouTube’s download protections to collect the creators’ videos for its AI video models. The court held that YouTube’s tokens plausibly “effectively control access” although the videos are free to watch, and that lost ad revenue is an injury traceable to the bypass itself. Snap lost the same point on 19 August, and Apple, before Judge Richard Seeborg, on 9 October. How the videos were gathered is still allegation.

Read

ChatGPT Is Eating the WorldChatGPT Is Eating the WorldMediaPost

Oct 1, 2026

The Justice Department Asks an Appeals Court to Let the Register of Copyrights Be Removed

In a 71-page brief to the D.C. Circuit in Perlmutter v. Blanche, the Justice Department asked the court to uphold the ruling that denied Shira Perlmutter an injunction — which would clear the way to remove her as Register of Copyrights. It argues the Librarian of Congress and the Register “exercise executive power and thus must be components of the Executive Branch”, and that she is now “exercising executive power outside the control and supervision of the President”. Perlmutter’s brief maintains the Library is a legislative agency. She remains in office under an injunction pending appeal granted in September 2025.

Read

Yahoo News

Sep 29, 2026

An Appeals Court Rules on AI Training for the First Time — and Says Its Case Is Not the Generative One

The Third Circuit affirmed Thomson Reuters’ win over ROSS Intelligence (No. 25-2153), the first time a federal appeals court has ruled on whether training an AI system on copyrighted material is fair use. It held 2,243 Westlaw headnotes copyrightable and ROSS’s copying of them not fair: the use was commercial and “minimally transformative, at best”; the underlying opinions were freely available — “Unlike necessity, ease is not a justification for copying” — and ROSS usurped a “rapidly developing” market for licensing headnotes as training text. The court limited itself by distinction, not reservation: “Unlike the AI models in Bartz and In re: OpenAI, ROSS’s AI platform cannot generate original expression.” It called the appeal “an ordinary copyright case”, noted the Justice Department had filed for OpenAI but “notably” not here, and cited no Copyright Office report.

Read

US Court of Appeals for the Third CircuitBallard Spahr (Consumer Finance Monitor)Music Business Worldwide

Sep 29, 2026

Brussels Opens a Consultation on Copyright and AI

The European Commission opened a targeted consultation on a possible copyright initiative, covering how protected work is used in AI — licensing, enforcement and AI imitations of performers among its sections. Responses are due by 3 November. It proposes no legislation and changes no existing obligation. In July the Commission published a study on whether to build an EU registry of the opt-outs by which rightsholders reserve their work from text and data mining under Article 4(3) of the 2019 copyright directive.

Read

Kluwer Copyright Blog

This quarter

Sep 18, 2026

Perlmutter’s Suit Is Paused to See Whether Congress Settles Who Appoints Her

Judge Timothy Kelly of the federal court in Washington stayed Shira Perlmutter’s suit over her firing as Register of Copyrights until 18 December, citing “the obvious effect on this case” if H.R. 6028 becomes law. The bill, which the House passed by voice vote on 8 June and which awaits the Senate, would make the Register an officer appointed by the President and confirmed by the Senate. If it is enacted, the parties must report within seven days. The order is quoted here from IPWatchdog’s account of it. Her appeal in the D.C. Circuit continues separately.

Read

IPWatchdog

Sep 17, 2026

An Unsealed Filing Shows How the Newspapers Say Their Articles Reached OpenAI

A less-redacted summary-judgment brief from news publishers (25-md-3143, Dkt. 1977-1) says how their work was gathered: OpenAI’s scrapes and Common Crawl downloads — one internal list, they say, shows 2,182,079 Chicago Tribune and 2,064,805 Times articles in a single training set — a 1.8-million-article Times archive licensed for non-commercial research, and “billions” of Bing-indexed pages passed from Microsoft to OpenAI. It quotes a Microsoft science director calling this “an astonishing theft of unprecedented proportions”; Microsoft says that was one employee’s view, not its own. All of it is the plaintiffs’ account of sealed evidence; the Tribune is one of them.

Read

CourtListener / RECAP document archive (Free Law Project)Chicago Tribune

Sep 16, 2026

The Ninth Circuit Closes the DMCA Route Against Copilot — and Leaves Training Undecided

The Ninth Circuit affirmed dismissal of programmers’ DMCA claims against GitHub, Microsoft and OpenAI (Doe v. GitHub, No. 24-7700). Section 1202(b) punishes removing copyright notices from copies of a work; on the programmers’ own account, the court held, Copilot and Codex “create new works that never contained that information”. It declined to make “identicality” a separate test, and warned of “potentially ruinous liability” if resemblance alone triggered the DMCA’s damages of up to $25,000 a violation. It decided nothing about training: that theory was forfeited below, and the court expressed “no view” on infringement. Two contract claims continue.

Read

Courthouse News ServiceAuthors Alliance

Sep 15, 2026

ANI Appeals, and Delhi Declines to Freeze the Ruling That Favoured OpenAI

A two-judge bench of the Delhi High Court, Justices Avneesh Jhingan and Manmeet Pritam Singh Arora, heard ANI’s appeal against the July order that refused to stop OpenAI training on its reporting. It declined to stay that order without hearing OpenAI, sought its response and set the appeal for 8 December. ANI argues India’s fair-dealing exception is closed and purpose-specific, that commercial use at scale cannot be “private or personal use, including research”, and that the judge in effect imported American fair use. Before suing, ANI had offered OpenAI a licence to its whole archive for $7.5 million.

Read

The Indian ExpressLiveLaw

Our reading

Sep 14, 2026

The Cases Are Splitting Into Three Questions, and the Plaintiffs Are Moving

An analysis of 105 classifiable US AI copyright filings — 89 litigation families — finds 77 per cent in just two courts: 49 in the Northern District of California, 32 in the Southern District of New York. Filings went from 8 in 2023 to 15, then 32, then 47 in 2026. What is inside them is shifting. Of the 47 dated 2026 filings, 33 allege training, 28 raise DMCA theories, and **14 allege unauthorised acquisition or piracy — against 6 of 32 the year before.** Courts have begun separating three questions early coverage treated as one: whether training is fair use, whether the material was lawfully obtained, and whether outputs infringe.

Compared

Legalytics (Adam Feldman)Justia (court documents)

Sep 8, 2026

The Fair-Use Question Reaches Summary Judgment

Competing summary-judgment motions in the New York Times’ case against OpenAI and Microsoft ask Judge Sidney Stein, in the Southern District of New York, to decide whether training on copyrighted work is fair use. The posture is the news: nearly every AI copyright ruling so far has been at the pleading stage, deciding only whether a claim could proceed. A summary-judgment ruling would decide the question itself. Reuters frames the two sides — authors and news outlets saying generative AI threatens writers, OpenAI and Microsoft calling the training transformative and saying it does not. No hearing or ruling date is available; the article is behind a paywall and only its headline and subheads could be read.

Read

Reuters

Sep 4, 2026

Both Sides Ask the Court to Decide Whether Training Is Fair Use

The largest AI copyright case reached the merits on 4 September, when both sides in the consolidated In re: OpenAI litigation (25-md-3143, S.D.N.Y.) moved for summary judgment before Judge Sidney Stein. OpenAI filed twice — a 46-page brief in the authors' class case (Authors Guild v. OpenAI, 23-cv-8292, Dkt. 1221) and a 33-page brief answering the news publishers (23-cv-11195, Dkt. 1496) — arguing that pretraining is fair use as a matter of law, that nobody owns the facts a newspaper reports, and that its web-browsing feature was impliedly licensed because the papers left its crawler unblocked for months. The Authors Guild and thirteen authors, among them George R.R. Martin, John Grisham and Jonathan Franzen, cross-moved on 194 works and argue Microsoft is vicariously liable. After a two-week extension, oppositions are due on 23 October and replies on 20 November; expert challenges are due on 29 October.

Read

CourtListener / RECAP document archive (Free Law Project)CourtListener / RECAP document archive (Free Law Project)Authors GuildChatGPT Is Eating the WorldAuthors GuildChatGPT Is Eating the World

Sep 4, 2026

OpenAI Concedes It Trained GPT-3 on a Pirate Library, and Asks the Court to Say It Does Not Matter

In its books brief OpenAI states plainly that compilations from Library Genesis — which the US Trade Representative lists as a notorious market for piracy — trained GPT-3 in spring 2020 and GPT-3.5 in late 2021, and that no other model in the case used it. It then asks the court to reject the line Judge Alsup drew in Bartz v. Anthropic, arguing that where the end purpose is transformative "the entire course of conduct undertaken in service of that transformative purpose—including intermediate copying—is fair use". The authors reply that Warhol requires each act of copying to be judged on its own.

Read

CourtListener / RECAP document archive (Free Law Project)Authors Guild

Sep 4, 2026

Searching a sample of 20 million ChatGPT conversation logs produced in the case, OpenAI's expert found 24 instances of verbatim regurgitation of the asserted news articles — a rate of 0.00012 percent — and the longest verbatim passages anyone identified ran to 29 and 43 words.

— OpenAI

Sep 4, 2026

Two More Newspapers Sue OpenAI, With a Claim the Times Had Just Given Up

The Seattle Times and Newsday filed a 38-page copyright and trademark complaint against OpenAI and Microsoft in the Southern District of New York, alleging the companies scraped their sites, paywalled articles included, to train and run ChatGPT, Copilot and Bing's AI features. They want damages and the "impoundment and/or destruction" of any dataset or model containing their work, and cite a 47 percent year-on-year fall in search referrals to midsize publishers in December 2025. Their trademark claim — over fabricated stories attributed to them — is the kind the Times abandoned with prejudice four weeks earlier. No docket number has been published.

Read

ReutersNewsdayThe Seattle Times

Sep 1, 2026

The Justice Department Tells the Court AI Training Is Fair Use — and That Its Own Copyright Office Was Wrong

The United States filed a 20-page Statement of Interest under 28 U.S.C. §517 in the consolidated In re: OpenAI Copyright Infringement Litigation (No. 25-md-3143, S.D.N.Y., Judge Sidney Stein), as the cases head into summary judgment — the first time the federal executive has taken a merits position on AI training. It urges the court to reject any argument that training infringes, calling the use "transformative — spectacularly so", and attacks the "market dilution" theory Judge Chhabria floated in Kadrey as "deeply flawed": under Warhol, outputs are a separate use from training, and works merely in the same genre are not market harm. Pointedly, it says the Register of Copyrights — "currently challenging her removal" — endorsed that theory in the Office's May 2025 report, that her view "does not warrant deference", and that her reasoning was "threadbare". It grounds the position in AI "dominance" and warns mandatory licensing would build an "oligopoly".

Read

CourtListener / RECAP document archive (Free Law Project)ChatGPT Is Eating the WorldTechCrunch

Aug 28, 2026

Sony and Warner Sue Anthropic — and Name Amodei and Mann Personally

Sony Music Publishing, Warner Chappell and about three dozen affiliated publishers filed a 48-page complaint in the Northern District of California, and the caption is the news: CEO Dario Amodei and co-founder Benjamin Mann are defendants in their personal capacities. Two counts run against the individuals — direct infringement from torrenting, and contributory infringement for the same acquisition conduct — and two against the company, including a DMCA §1202 claim for stripping copyright management information, which carries $2,500 to $25,000 per violation on its own. The publishers allege lyrics were taken from Books3, The Pile, Common Crawl and the aggregators MusixMatch and LyricFind, covering tens of thousands of compositions, and that Claude reproduces them verbatim. Anthropic calls it "the third lawsuit from the same lawyers, recycling allegations from cases already before the courts" and says training is fair use. The complaint itself was not obtained.

Read

The IndependentSigma Law Group

Aug 21, 2026

WikiHow Sues OpenAI, and Asks to Be Folded Into the MDL

wikiHow, Inc. filed a copyright infringement action against nine OpenAI entities in the Southern District of New York on 21 August — wikiHow, Inc. v. OpenAI, Inc. et al, No. 1:26-cv-07171, cause 17:101 — represented by Susman Godfrey. With it came a statement of relatedness asking that the case be treated as related to the consolidated OpenAI copyright proceeding, MDL No. 1:25-md-03143, which is the structurally interesting part: the docket is consolidating rather than fragmenting. The complaint itself could not be read. Reporting describes unauthorised scraping of more than 11,000 how-to guides and verbatim reproduction by ChatGPT; that allegation is carried as reported and has not been checked against the pleading.

Read

PacerMonitorABAB News

Aug 20, 2026

An AI Training Claim Is Dismissed for Not Saying Where the Books Came From

Judge Jacqueline Scott Corley dismissed the part of author E. Molly Tanzer's claim against Adobe that concerns NVIDIA's Nemotron model, in Tanzer v. Adobe Inc., No. 3:26-cv-04712-JSC (N.D. Cal., 20 August). The ruling decides nothing about fair use. It is about pleading: alleging that NVIDIA "acquired millions of unauthorized copies" from the shadow library Anna's Archive, and that piracy "has been commonplace with the AI development community", is conclusory — and, the court wrote, repeating a conclusory allegation several times does not make it less conclusory. Without facts supporting an inference that the model saw her books by that route there is neither Article III standing nor a direct infringement claim. Dismissal is with leave to amend by 2 October. The obstacle in front of plaintiffs is increasingly provenance, not the exception.

Read

Wolters Kluwer CCH IP Law Daily (court document archive)

Aug 17, 2026

A music publisher sues Suno and Anthropic for a billion each — and names the scraper that fed them

Round Hill Music LP and its royalty funds filed two copyright complaints in the US District Court for the Northern District of California, against Suno (reported as No. 5:26-cv-08507) and against Anthropic (No. 5:26-cv-08505). Round Hill manages a US$1.1 billion portfolio of more than 200,000 songs — among the asserted works are the Goo Goo Dolls' "Iris", Bonnie Tyler's "Total Eclipse of the Heart" and James Brown's "I Got You (I Feel Good)" — and is represented by Richard S. Busch, who won the "Blurred Lines" verdict for the Marvin Gaye family. Each complaint pleads 500 "bell-weather" compositions preliminarily and states that the plaintiffs intend to amend to list "potentially ten thousand or more" works, so that statutory damages "would ultimately exceed hundreds of millions of dollars and potentially approach or conceivably exceed $1 billion". Three things make the filings more than another entry on the docket. They plead the Digital Millennium Copyright Act alongside the Copyright Act, alleging both the removal of copyright management information and the circumvention of technological access controls to pull protected audio from licensed platforms. They take Judge Alsup's line in Bartz — that a permanent library of unlicensed copies is not fair use — and turn it against training itself, arguing that a "permanent stockpile of copyrighted material for open-ended future exploitation fails every factor of the fair-use analysis" because "the purpose is acquisitive rather than transformative". And the Suno action names a third defendant that is not a model builder at all: Bright Data, the Israeli web-scraping vendor, on a contributory theory that its tools are "specifically designed to scrape music files en masse and strip them of CMI", pleading that it passed $300 million in revenue in 2025 with $400 million projected for 2026. On memorisation, the Anthropic complaint alleges that "despite implementing partial output filters in response to prior litigation", Claude still reproduces "the vast majority of Round Hill's catalog". These are allegations; no court has tested any of them.

Read

CourtListener / Free Law ProjectMorningstarDigital Music NewsThe Hollywood Reporter

Aug 17, 2026

Round Hill is not against artificial intelligence but against building billion-dollar businesses on other people's creative work while paying the creators nothing; licensing protects the legal owners of the raw material, and the company intends to take both cases to trial rather than accept a resolution that leaves songwriters and artists without their share.

— Josh Gruss

Aug 13, 2026

Two copyright suits against Suno end without a court reaching the merits

S.A. Jamendo, the Luxembourg subsidiary of the Winamp Group, filed a voluntary notice of dismissal without prejudice in the US District Court for the District of Massachusetts (No. 1:26-cv-12966-AK), six weeks after suing Suno over the MTG-Jamendo Dataset, which draws more than 55,000 tracks from its catalogue and was released for non-commercial research on terms requiring a licence for commercial AI training. The notice, signed by attorney Alexander T. Hornat, gives no reason and mentions no settlement, and records that no defendant had served an answer or a motion for summary judgment — leaving Jamendo free to bring the same claims again. Around the same time Poseidon Wave Media, the entity behind the instrumental duo The American Dollar, dismissed its own suit against Suno with prejudice, so it cannot be refiled; it had been filed in the Southern District of New York on 12 May 2026 and asserted 236 recordings and compositions on a claim that the duo's licensing revenue had been nearly eliminated. Jamendo's parallel suit against Nvidia over the same dataset, in the Northern District of California, has not been withdrawn, and a related commercial claim of about €16 million against Nvidia Technologies Belgium is pending in Belgium's Ghent Enterprise Court. Neither dismissal states a reason, and neither is reported as settled; Digital Music News suggests Jamendo, headquartered in Luxembourg, may prefer a European forum after GEMA's Munich win, which is that outlet's inference and not a finding of the court. The larger actions continue: Universal Music Group and Sony Music Entertainment are still suing Suno in the same Massachusetts district.

Read

Music Business WorldwideMusic Business WorldwideDigital Music NewsCourtListener / Free Law Project

Aug 12, 2026

Suno buys its way out of the German ruling's logic, licensing BMG's catalogue twelve days later

BMG and Suno announced a global strategic alliance covering BMG's recorded music and publishing repertoire, twelve days after the Regional Court of Munich I held that Suno was not entitled to train on, store or reproduce works represented by the German rights society GEMA. The agreement is built on an opt-in principle — BMG artists and songwriters who choose to participate have their rights protected and are compensated — and it also settles Suno's prior use of BMG recordings and publishing works, which is the part that matters to the price of a work: an unlicensed corpus is retrospectively paid for rather than adjudicated. No financial terms were disclosed by either party. It follows Suno's November 2025 settlement and licensing partnership with Warner Music Group, and comes six days after Suno's co-founder and chief executive Mikey Shulman published company principles committing to audio watermarking and fingerprinting and to a new downloads policy limiting mass distribution of generated tracks. BMG is simultaneously a plaintiff against Anthropic in the lyrics litigation, so the same company is buying a licence from one model builder while suing another.

Read

BMGMusic Business WorldwideSuno

Aug 6, 2026

A Supreme Court Ruling Strips Three Theories Out of the New York Times Case

A month before the summary judgment motions, Judge Sidney Stein recorded the damage a new Supreme Court decision had done to the news publishers' case. In Cox Communications v. Sony Music (146 S. Ct. 959 (2026)) the Court held that "material contribution" is not a valid theory of contributory copyright liability — the theory on which The New York Times, the Daily News and the Center for Investigative Reporting had all pleaded that OpenAI and Microsoft were answerable for their users' infringement. Facing OpenAI's motion for judgment on the pleadings, all three dropped those claims with prejudice; the Times and Daily News dropped their trademark dilution claims too. Stein then refused to let them re-plead inducement against Microsoft, holding those theories existed before Cox and switching now "would require reopening discovery into Microsoft's intent".

Read

CourtListener / RECAP document archive (Free Law Project)

Jul 31, 2026

Munich rules for GEMA against Suno: training on protected music is itself the infringement

The Regional Court of Munich I (Landgericht München I, 42nd Civil Chamber, presiding judge Elke Schwager) ruled for the German rights society GEMA against the AI music generator Suno in first-instance judgment 42 O 763/25, holding that Suno infringed copyright by training on six protected works without a licence and by reproducing them in its outputs. The court found the model had ‘memorised’ and stored the works in reproducible form, which defeats the text-and-data-mining exception; held the provider rather than the user liable; held compliance with the EU AI Act no defence; and found the U.S. training copies not protected by fair use, expressly distinguishing the American Bartz and Kadrey rulings on the ground that there the training materials were not reproduced in outputs. It took jurisdiction over the U.S. training acts under the German Collecting Societies Act (VGG), noting Suno had used stream-ripping to pull tracks from YouTube in circumvention of its ‘rolling cipher’ protection. It ordered Suno to cease the reproduction and the use of the works for training, to disclose the revenue connected to the infringements before damages are set, and confirmed an entitlement to damages whose amount is not yet fixed. It is the second time the same court has held that AI training infringes German copyright — it held in November 2025 in GEMA v OpenAI that training on song lyrics infringes — and it moves German law further from the American line in Bartz v. Anthropic, where training on lawfully acquired works was fair use and only pirated acquisition was punished: under the Munich rulings the training itself is the infringement, and licensing, not acquisition hygiene, is the price of entry. The judgment is first-instance and not final; Suno intends to appeal. The disclosure order lands on a company carrying a $5.4bn valuation and roughly $300m in annual recurring revenue, while parallel U.S. suits by Universal and Sony seek up to $150,000 per work. Accounts differ on immediate enforceability — juve-patent reports the judgment is not yet enforceable, while Forbes reports a German first-instance judgment can be enforced pending appeal, typically on posting security.

Read

Bavarian State Ministry of JusticeJUVE PatentReed SmithGEMAForbesBird & BirdVariety

Jul 31, 2026

AI models built on stolen intellectual property have no protection under the law, and providers must pay for licenses.

— Tobias Holzmüller Mixed

Read

BMG

Earlier this year

Jul 24, 2026Delhi's High Court refuses to stop OpenAI, reading Indian fair dealing to cover model training
Jul 22, 2026Music publishers escalate the Anthropic lyrics fight with a discovery-fueled complaint
Jul 22, 2026“By contrast, Anthropic's copyright infringement is not innovation. In layman's terms, it's theft.” — Concord Music Group
Jul 21, 2026Court grants final approval to Anthropic's $1.5 billion settlement
Jul 21, 2026“We applaud the court's final approval of this settlement, which represents an important victory in the larger battle to hold big tech accountable for its unscrupulous appropriation of intellectual and creative properties that clearly belong to authors and publishers.” — Maria Pallante±Fact check: mixed
Jun 30, 2026Supreme Court leaves the ousted Register of Copyrights in office — for now
Nov 13, 2025Cohere Loses Its Motion to Dismiss, Entirely
Nov 11, 2025A Munich court holds that training on song lyrics infringes German copyright
Nov 4, 2025UK High Court largely clears Stability AI after Getty drops its core claims
Nov 4, 2025“The ruling resolves the core copyright concerns in the case, and Getty voluntarily dismissed most of its copyright claims at the close of trial testimony.” — Stability AI±Fact check: mixed
Nov 4, 2025“Even well-resourced companies face significant challenges protecting creative works given the lack of transparency requirements, and governments including the UK should establish stronger transparency rules.” — Getty Images

Earlier

Sep 5, 2025
Anthropic settles the pirated-books class action for $1.5 billionAnthropic agreed to pay about $1.5 billion to settle the Bartz class action over pirated books used to train Claude — the largest copyright recovery on record and the first major AI-era settlement. It covers an estimated 500,000 works at roughly $3,000 each, with Anthropic also agreeing to destroy the downloaded files, resolving the piracy liability that survived Alsup's June ruling and averting the December trial. Court filings had shown Anthropic downloaded more than 7 million books it knew were pirated (from Books3, LibGen and the Pirate Library Mirror) before shifting to buying and scanning physical books.
Read
Associated Pressapnews.com95% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.Authors Guildauthorsguild.org58% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.
Sep 5, 2025“The Anthropic settlement will do little for European writers whose works are not registered with the US Copyright Office, and fits a tech-industry playbook of growing first and later paying a relatively small fine for breaking the rules.” — Danish Rights Alliance
Aug 14, 2025Getty drops its US case against Stability AI to refile in California
Jun 25, 2025A second judge reaches the opposite instinct in Kadrey v. Meta
Jun 25, 2025“In most cases it will likely be illegal to copy copyrighted works to train generative AI without permission, because AI trained on those works can generate countless competing works and thereby dilute the market for the originals — so companies will generally need to pay for training rights.” — Vince Chhabria
Jun 23, 2025Alsup rules AI training fair use — but piracy is not
Jun 6, 2025Ousted Register of Copyrights sues the president
May 10, 2025Register of Copyrights fired a day after the Part 3 report
May 9, 2025Copyright Office Part 3 floats a "market dilution" theory of AI-training harm
Feb 20, 2025The DMCA Claims Keep Failing Where the Copyright Claims Survive
Feb 11, 2025First US AI fair-use ruling goes against the AI company
Feb 11, 2025“The ROSS court characterized Westlaw headnotes as original works — analogizing an editor's distillation of a judicial opinion to a sculptor's carving of a figure from raw stone — and distinguished them from the uncopyrightable opinions they summarize.” — Stephanos Bibas✓Fact check: true
Jan 29, 2025US Copyright Office issues Part 2 on the copyrightability of AI output
Jan 3, 2025Anthropic agrees to lyric "guardrails" in the music-publishers suit
Aug 12, 2024Andersen v. Stability AI survives dismissal and heads to discovery
Jul 31, 2024US Copyright Office begins its landmark AI study with Part 1
Dec 27, 2023The New York Times sues OpenAI and Microsoft
Dec 27, 2023“OpenAI said it was "surprised and disappointed" by the New York Times lawsuit, saying it had been in constructive talks with the paper and respects the rights of content creators.” — OpenAI
Dec 27, 2023“A Supreme Court decision on AI training and copyright is essentially inevitable, with some publishers settling but enough litigating to force resolution of the novel question.” — Richard Tofel
Sep 7, 2023Microsoft pledges to indemnify Copilot customers against copyright suits
Feb 3, 2023Getty Images sues Stability AI over 12 million photographs
Jan 12, 2023Artists file the first class action against AI image generators

Covered since Jan 2023Read it from the start →