Menu
Home
Forums
New posts
Search forums
What's new
Featured content
New posts
New media
New media comments
New resources
Latest activity
Media
New media
New comments
Search media
Resources
Latest reviews
Search resources
Nyuuz
Log in
Register
What's new
Search
Search
Search titles only
By:
New posts
Search forums
Menu
Log in
Register
Install the app
Install
Home
Forums
Labrish
Nalij
Jinaral kantent
Missing song metadata creates a tougher DMCA claim
JavaScript is disabled. For a better experience, please enable JavaScript in your browser before proceeding.
You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an
alternative browser
.
Reply to thread
Message
[QUOTE="Bombastus, post: 91335, member: 2178"] Gerencia 360’s August 31 complaint accuses Suno and Bright Data of intentionally stripping identifying information from 1,174 copyrighted works. The allegation sits under Section 1202 of the DMCA, a provision aimed at the integrity of copyright management information rather than access controls. That makes it legally separate from the lawsuit’s claims about paywalls, CAPTCHAs, and other barriers. Copyright management information can include a work’s title, author, copyright owner, performer, usage terms, and identifying numbers connected to that information. Gerencia says its songs and recordings carried details such as titles, songwriter names, publisher information, copyright notices, and ISRC codes when displayed on licensed platforms. The complaint alleges those details disappeared during scraping and data preparation. Missing credits alone do not establish liability. Section 1202 asks whether information was intentionally removed or altered without authority and whether the defendant knew, or had reasonable grounds to know, that the conduct would induce, enable, facilitate, or conceal infringement. That second requirement makes [B]deliberate metadata stripping[/B] much harder to prove than simply showing a cleaned training file. [HEADING=2]Section 1202 demands more than missing credits[/HEADING] The Ninth Circuit put real weight on that knowledge requirement in Stevens v. CoreLogic. Professional photographers complained that software processing their images removed embedded copyright metadata, but the court held that they still needed evidence connecting that removal to likely future infringement. An automated process that drops metadata is not enough by itself. The court wanted an affirmative showing that the defendant understood the probable consequences of removing the information. Evidence of a pattern, a known workflow, or a method designed to make infringement easier can matter. The standard is sometimes described as double scienter because knowledge attaches both to the prohibited conduct and to its likely relationship with infringement. That creates a sharper test for [B][URL='https://goldmidi.com/community/threads/gerencia-360-sues-suno-bright-data-in-california.76732/unread']Gerencia 360’s copyright metadata allegations[/URL][/B] than the damages headline suggests. Gerencia alleges that engineers selected a data-extraction method that aggressively removed authorship and ownership information despite alternatives that could preserve it. It also claims Bright Data’s extraction tools discarded identifying material while assembling scraped datasets for AI developers. Those allegations remain unproved. Still, they try to answer the exact weakness that has killed other Section 1202 claims, where plaintiffs could show missing metadata but could not plausibly connect the loss to a deliberate infringement-related purpose. [HEADING=2]Anthropic showed how a stronger pleading can survive[/HEADING] A 2025 Northern District of California ruling in the music publishers’ case against Anthropic offers a useful comparison. The court had previously dismissed the publishers’ copyright-management claim, then allowed an amended version to proceed after they added more specific allegations about Anthropic’s data-processing choices. The amended complaint alleged that Anthropic deliberately selected a particular text-extraction algorithm that removed more copyright information than available alternatives. More importantly, the publishers alleged that the choice was intended to prevent the model from returning identifying information when users prompted it for copyrighted material. At the pleading stage, that was enough to support the required knowledge inference. Gerencia uses a strikingly similar structure. Its complaint says Suno knew models can memorize and reproduce training material, and that removing attribution during training would help ensure reproduced lyrics appeared without identifying their original authors. It alleges that this could conceal Suno’s own copying while making later infringement by users harder to recognize. That does not guarantee the claim survives discovery or trial. Internal engineering records could show a purposeful stripping process, an ordinary normalization step, or something between those extremes. The legal difference can turn on why a processing choice was made and what the people selecting it understood about its downstream effects. [HEADING=2]Suno’s outputs may create the harder second fight[/HEADING] Section 1202 also reaches certain distributions of works when copyright management information has already been removed. Gerencia therefore alleges not only stripping during the training pipeline but distribution through Suno outputs without the original identifying information. That theory brings the relationship between the source work and generated output into focus. Courts have disagreed over whether Section 1202 requires the distributed copy to be identical to the original work from which the information was removed. A Northern District of California judge adopted an identicality requirement in the GitHub Copilot litigation, while other courts have rejected such a strict reading. As of September 3, 2026, that issue remains pending before the Ninth Circuit after oral argument in February. The distinction matters for AI music. A verbatim lyric reproduction with the songwriter and publisher information stripped presents a cleaner factual theory than a newly generated passage that merely resembles a protected song. If an output substantially transforms the source material, defendants can argue that no copy exists from which copyright management information was actually removed. Gerencia’s strongest Section 1202 evidence may therefore come from the pipeline rather than from every generated song. Records showing that identifiable source files entered a cleaning process with ownership information attached and emerged without it could establish a concrete removal event. The remaining fight would be whether that removal was intentional and whether the defendants understood that it would make infringement easier to carry out or conceal. [/QUOTE]
Insert quotes…
Name
Post reply
Home
Forums
Labrish
Nalij
Jinaral kantent
Missing song metadata creates a tougher DMCA claim
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
Accept
Learn more…
Top