copyright – internetcases – Attorney Evan Brown

April 16, 2026 Artificial Intelligence, DMCA

AI model training case raising DMCA anticircumvention issues moves forward

DMCA anticircumvention for training AI models

Plaintiff record companies sued defendant Udio, an AI music generation platform that allows users to create songs from text prompts and audio inputs. Plaintiffs alleged violations of the Digital Millennium Copyright Act based on assertions that defendant – by using stream ripping tools such as YT-DLP – circumvented technological protections to copy copyrighted music from YouTube and use that music to train its models.

Defendant asked the court to dismiss the DMCA anticircumvention claim, arguing that plaintiffs failed to plausibly allege that YouTube’s technological measures qualified as protected access controls under the statute or that defendant’s conduct constituted unlawful circumvention.

The court denied the motion. It held that plaintiffs plausibly alleged that YouTube employs technological measures that control access to copyrighted works and that defendant circumvented those measures, and the court held that the technical classification of those measures could not be resolved without a more developed factual record.

Sony Music Entertainment v. Uncharted Labs, Inc., 2026 WL 1019199 (S.D.N.Y. Apr. 15, 2026)

April 6, 2026 Copyright

Teacher’s X repost was fair use and not copyright infringement

The United States District Court for the Southern District of Indiana dismissed with prejudice all claims brought by plaintiff author Keith Bell against Bartholomew Consolidated School Corporation and one of its teachers and allowed them to seek attorneys’ fees. Bell, the author of the book Winning Isn’t Normal and the owner of the copyright in the work, filed suit after a March 2022 social media repost involving a passage from his work.

The parties

The dispute centered on the conduct of one Timothy Bless, a high school teacher and football coach employed by the school system. Bless was not acting on behalf of the school in any official capacity when the events occurred. Instead, he used his personal X account, which he controlled independently. On that account, Bless reposted a post originally made by a college football coach. The post contained an image of one of Bell’s well-known motivational passages. Bless did not add any commentary or attempt to connect the repost to the school or any official activity.

Bell alleged that this repost violated copyright law and a prior settlement agreement between the parties. He also brought claims for vicarious liability against the school and for violations of the Digital Millennium Copyright Act. He sought damages and a finding that the defendants had used his copyrighted work without authorization.

The court’s decision – fair use

The court rejected all of Bell’s claims. The central issue was whether Bless’s repost constituted copyright infringement. The court concluded that it did not because the repost qualified as fair use. In reaching this conclusion, the court focused heavily on the purpose and context of Bless’s actions. It found that Bless shared the passage to motivate or inspire student athletes, not for any commercial purpose. There was no evidence that he profited from the repost or used it in connection with any business activity. This noncommercial and educational context weighed strongly in favor of fair use.

The court also considered the nature of the work and the amount used. While the passage had some creative elements, it largely consisted of familiar motivational ideas. Although Bless reposted the entire passage, the court found that copying the whole excerpt was reasonable because its full text was necessary to convey its inspirational message. The fact that the passage was already widely available online further supported this conclusion.

Most importantly, the court found no plausible harm to the market for Bell’s work. There was no indication that the repost reduced sales of the book or related merchandise. If anything, the court suggested that widespread sharing of the passage could increase interest in Bell’s work rather than diminish it. Because market harm is often the most significant factor in fair use analysis, this finding strongly supported the court’s conclusion.

Other claims failed as well

The claims against the school also failed because Bell did not plausibly allege that the school itself copied the work or controlled Bless’s personal social media activity. The repost occurred on Bless’s private account, and there were no facts linking the school to that decision. The vicarious liability claim was dismissed as conclusory, and the DMCA claim failed because Bell did not adequately allege that copyright information had been knowingly removed. The breach of contract claim was also dismissed because Bell did not plausibly allege actionable use or damages.

Given Bell’s litigation history and the weakness of the claims, the court determined that any amendment would be futile and dismissed the case with prejudice. It also concluded that an award of attorneys’ fees was appropriate, in part to deter similar claims in the future.

Bell v. Bartholomew Consolidated School Corporation, 2026 WL 828677 (S.D. Ind. Mar. 26, 2026).

March 25, 2026 Copyright, DMCA

Supreme Court rejects contributory liability in long-running Cox case

Plaintiffs sued Defendant for secondary copyright infringement, alleging that Defendant, an internet service provider, was liable because it continued to provide internet access to subscribers whose accounts were associated with music piracy. Plaintiffs won a jury verdict for $1 billion, and the Fourth Circuit Court of Appeals agreed in part, concluding that Defendant could be contributorily liable for continuing to serve known infringers.

Defendant asked the U.S. Supreme Court to reverse the contributory infringement ruling, arguing that merely providing internet service to subscribers suspected of infringement does not make a service provider liable under the Copyright Act. It contended that secondary liability requires proof that Defendant induced infringement or provided a service designed for infringement, not just knowledge that some subscribers used the service unlawfully.

The Court ruled that Defendant was not contributorily liable and reversed the Fourth Circuit’s decision on that issue. It held that supplying internet access to the general public, even with knowledge that some users may infringe copyrights, is not enough by itself to establish contributory copyright infringement.

The Court ruled this way because contributory liability requires intent, which Plaintiffs could show only by proving that Defendant induced infringement or provided a service tailored to infringement. The Court concluded that Defendant did neither, since it did not encourage infringement and its internet service had substantial lawful uses. The Court also rejected Plaintiffs’ reliance on the DMCA, explaining that the statute creates safe harbors and does not itself impose liability on service providers that fail to qualify for them.

Cox Communications, Inc. v. Sony Music Entertainment, 607 U.S. ___, 2026 WL 815823 (Mar. 25, 2026).

January 12, 2026 Artificial Intelligence, Copyright, Litigation

Infringement case against OpenAI failed because there was no copyright registration

Thinking about suing an AI company for copyright infringement? Do not overlook the basics. Before any court will consider the merits of an infringement claim, the plaintiff needs to have an actual copyright registration in hand, not just a pending application.

That notion was confirmed in a recent unsuccessful lawsuit against OpenAI in federal court in California. Plaintiff sued OpenAI alleging that OpenAI infringed the copyright in several artificial intelligence models and content plaintiff claimed to have developed and then destroyed evidence to conceal that alleged infringement.

Plaintiff asked the court to issue a temporary restraining order preventing defendant from deleting or altering data and documents related to the alleged infringement while the case proceeded. The court denied the request for a temporary restraining order and dismissed the complaint.

The court ruled this way because the Copyright Act bars any civil infringement action until copyright registration has been made, and courts interpret that requirement to mean the Copyright Office must have issued a registration certificate, not merely received an application. This left plaintiff in this case unable to show a likelihood of success on the merits.

Gholami v. OpenAI, Inc., No. 26-cv-00174, 2026 WL 61359 (N.D. Cal., January 8, 2026).

Here is some relevant info:

January 6, 2026 Copyright, DMCA, Social Media

Does the DMCA safe harbor cover infringing images in an email?

Plaintiff photographer sued Pinterest for copyright infringement, alleging Pinterest displayed his and other photographers’ copyrighted images in notifications sent outside of the Pinterest website. Pinterest moved for summary judgment, arguing it was protected under the safe harbor provisions of Section 512(c) of the Digital Millennium Copyright Act (“DMCA”). The court granted Pinterest’s motion and dismissed the case.

Pinterest is a familiar and massive social media platform, where individuals upload and share image-based “Pins” that function as visual bookmarks. The platform displays Pins in personalized feeds curated by algorithms and which contain advertisements labeled as “promoted.” Pinterest also delivers through notifications such as emails, in-app alerts, and push notifications, which contain hyperlinks that trigger display of images hosted on its servers. One such notification that plaintiff received included his copyrighted photograph, prompting him to file suit six days later.

The court found that Pinterest’s actions fell within the DMCA’s Section 512(c) safe harbor, which shields service providers from copyright liability for content stored at the direction of users. Because Pinterest raised this as an affirmative defense, it had the burden to prove every element of the safe harbor criteria, and the court concluded it had met both the statutory threshold and all required conditions.

Statutory threshold requirements under the DMCA

To qualify for the DMCA safe harbor, Pinterest had to meet several threshold statutory requirements that are found in Sections 512(c) and (i): it had to be a service provider, maintain a designated agent, implement a repeat infringer policy, and accommodate standard technical measures. The court found that Pinterest satisfied all four. As “one of the largest social media platforms in the world,” it operated a qualifying online platform as defined by the statute. The evidence showed that Pinterest maintained a registered agent with the Copyright Office and that it enforced a strike-based policy for repeat infringers. And the court found that Pinterest did not interfere with any recognized standard technical measures that plaintiff implemented with his works. (Plaintiff had asserted that he embedded certain metadata in his photographs, but he did not argue that this metadata qualified as a “standard technical measure” under the DMCA, nor did he claim that Pinterest interfered with it — in fact, he alleged that Pinterest preserved the metadata on its servers.)

How Pinterest met the required conditions

After finding that Pinterest satisfied the DMCA’s threshold requirements, the court turned to whether Pinterest’s conduct of sending out copyright protected images in off-platform notifications was protected under Section 512(c). To do so, Pinterest had to show three things:

the alleged infringement occurred due to user-directed storage;
Pinterest lacked actual or red flag knowledge of the infringement; and
Pinterest either had no right and ability to control the activity or did not receive a direct financial benefit from it.

The court evaluated each element in turn.

By reason of storage at the direction of a user

The court concluded that Pinterest met the first requirement for DMCA safe harbor protection: the alleged infringement occurred “by reason of the storage at the direction of a user.” It emphasized that the image at issue was not embedded in the notification itself but was instead hosted on Pinterest’s servers and accessed via a hyperlink contained in the notification. When a user opened the message, their software triggered a request to Pinterest’s server to retrieve and display the image, just as it would when accessing content directly through the platform. Because this method merely facilitated access to user-uploaded content without altering it, the court found the display was within the statutory definition.

No knowledge of infringement

The court found that Pinterest satisfied the second requirement for DMCA safe harbor protection by showing it lacked actual or red flag knowledge of the alleged infringement. Critically, Harrington never sent Pinterest a DMCA takedown notice or otherwise identified the allegedly infringing material before filing suit. The DMCA operates on a notice and takedown system: platforms are not required to proactively monitor user content but must respond once they receive proper notice. Because Harrington gave no such notice and offered no evidence that Pinterest otherwise knew about the specific image at issue, the court concluded there was no genuine dispute as to Pinterest’s lack of knowledge.

Control and financial benefit

The court found that Pinterest met the third and final requirement for DMCA safe harbor by showing it neither had the right and ability to control the alleged infringement nor received a financial benefit directly attributable to it. While Pinterest used algorithms to curate content and monetize its platform generally, the court held that this did not amount to the kind of “substantial influence” over user activity that would disqualify it under the DMCA. Pinterest did not direct users to upload specific content, nor did it participate in any purposeful conduct related to the display of plaintiff’s photo.

The court also rejected plaintiff’s claim that Pinterest profited directly from the infringement. Pinterest presented evidence that its notifications did not contain advertisements and that it earned no revenue specifically tied to the image in question. Plaintiff’s counter-evidence failed to show otherwise. Even if ads had appeared near the image, the law requires a direct connection between the infringing display and revenue, which was absent here. Therefore, Pinterest satisfied this final element of the DMCA safe harbor defense.

Harrington et al. v. Pinterest, Inc., No. 20-CV-5290, 2026 WL 25880 (N.D. Cal., January 5, 2026)

July 11, 2025 Artificial Intelligence, Copyright

Voice cloning case presents novel AI copyright issues

digital voice clone

Two professional voice actors based sued Lovo, a company that sells AI-generated voiceover software alleging, among other things, copyright infringement. Plaintiffs claimed that defendant took their voices without permission and used them to create digital clones. Those clones were sold to customers under the pseudonyms “Kyle Snow” and “Sally Coleman.” Defendant moved to dismiss the copyright claims. The court provided a mixed ruling.

How it started

Plaintiffs asserted that defendant first contacted them on the freelance platform Fiverr in 2019 and 2020. Defendant’s representatives allegedly assured plaintiffs the recordings would only be used for private research, not public or commercial projects. Based on those promises, plaintiffs delivered the recordings. Each was paid a few hundred dollars.

Several years later, plaintiffs discovered their voices had been cloned. In 2023, they heard a podcast produced using Lovo’s software. Plaintiff Lehrman’s friends and colleagues said the cloned voice sounded exactly like him. After further research, plaintiffs learned that defendant had built digital voice profiles from their recordings and offered those profiles to customers as part of a paid AI service.

What plaintiffs claimed

Plaintiffs made four different types of copyright claims. One was that defendant had directly copied Plaintiff Sage’s recording and used it in a promotional video for investors. Another was that defendant had used both of their recordings to train defendant’s AI system called Genny. They also claimed that the cloned voices produced by Genny were infringing. Finally, they argued that defendant should be liable for contributory infringement because it let its customers use those clones.

What the court decided

The court allowed one copyright claim to move forward. Plaintiff Sage alleged that defendant used a real recording of her voice in public investor presentations and YouTube videos. Defendant admitted to using a portion of the recording, and the court agreed that this could be outside the scope of the license.

The other copyright claims were not as successful. The court dismissed the copyright claim based on use of the voice recordings to train the AI model. Plaintiffs did not explain in their complaint how the training process worked or how it infringed their rights. The court said plaintiffs could amend the complaint and try again.

The court also dismissed the claim based on the AI-generated voices themselves. It explained that copyright law does not protect imitations of a voice, only the actual recordings. Plaintiffs had not alleged that the AI clones were identical to the originals, only that they sounded very similar. That kind of mimicry was not covered by copyright law.

Finally, the court dismissed the contributory infringement claim. Since there was no direct infringement based on the AI-generated outputs, the court said there could be no indirect liability either.

Lehrman v. Lovo Inc., — F.Supp.3d —, 2025 WL 1902547 (S.D.N.Y., July 10, 2025)

June 26, 2025 Copyright, Litigation

Court lets authors expand copyright case to target Databricks’ new AI models

amending complaing

Five copyright holders sued Databricks and Mosaic ML, claiming their copyrighted works were used to train artificial intelligence systems without permission. Plaintiffs originally alleged that Mosaic ML directly infringed their works by training its MPT large language models on datasets that included their works. Plaintiffs also accused Databricks, Mosaic ML’s parent company, of vicarious liability for that conduct.

After Databricks released a new set of AI models called DBRX, plaintiffs moved to amend the complaint. Plaintiffs asked the court to allow a new claim of direct copyright infringement against Databricks for allegedly using the same protected works to train DBRX. Plaintiffs also sought to update the list of copyrighted works allegedly copied. Defendants opposed the request, arguing that the amendment came too late and would unfairly change the case.

Timing

The court acknowledged that plaintiffs waited more than a year after DBRX was released before requesting to amend the complaint. That delay was significant, and plaintiffs did not provide a strong explanation. However, the court noted that discovery was still open, and key deadlines had not yet passed. Because the case was still active, the court said the delay alone was not enough to deny the motion.

Intent

Defendants claimed plaintiffs acted in bad faith by dragging out the case and making vague statements in court filings. But the court saw no signs of deliberate delay or dishonesty. Instead, it found that plaintiffs’ motion to amend reflected an effort to match the complaint with new information obtained through discovery.

Prejudice

Defendants argued that allowing new claims about DBRX would cause unfair prejudice by drastically changing the case. The court disagreed. It found that the parties were already engaged in discovery related to DBRX and that any added burden would be limited. Since the DBRX and MPT models might rely on overlapping data, the new claims would not require a completely new approach to the case.

Futility

Defendants also said the new claims were too vague and would not survive a challenge. But the court said such issues should be dealt with after the complaint is amended. Unless the new claims are clearly invalid, courts usually allow amendments and address legal sufficiency later in the process.

So the court granted plaintiffs’ motion to amend. The lawsuit will now include direct copyright infringement claims against Databricks based on its newer DBRX models, along with an updated list of works that plaintiffs claims were copied.

In re Mosaic LLM Litigation, 2025 WL 1755650 (N.D. California, June 25, 2025)

TikTok and Meta terms granted other users remix rights

Plaintiff sued TikTok and Meta after other users on those platforms incorporated clips from her video into their own posts, allegedly without her permission. She claimed this was copyright infringement and also alleged that TikTok failed to protect her from harassment by users in the comments of her live videos. Plaintiff filed the lawsuit on her own, without a lawyer.

Plaintiff asked the court to hold TikTok and Meta liable for copyright infringement and to consider tort claims against TikTok for harassment. But both companies responded by asking the court to dismiss the case. They pointed to the user agreements Plaintiff had accepted when she signed up. Those terms gave the platforms and their users broad rights to use, modify, and distribute any content she uploaded. TikTok also invoked immunity under 47 U.S.C. 230, a provision in federal law protecting platforms from liability for user-generated content.

The court agreed with the platforms. It found that plaintiff had granted TikTok and Meta valid licenses to use her video, so there could be no copyright violation. The court also ruled that it had no authority to hear the tort claims because plaintiff had not shown that the court had jurisdiction over those parts of the case. The court rejected plaintiff’s arguments that she did not fully understand the contracts or that the agreements were unfair. On appeal, the Tenth Circuit upheld the decision, finding no clear error in how the lower court handled the case and ruling that plaintiff had waived her right to challenge the licensing issue by not objecting to it specifically.

In the end, the court dismissed all claims against both companies. The court also declined to take up any new claims plaintiff tried to raise during the appeal, saying she had not brought those up earlier and did not support them with enough detail.

Three reasons why this case matters:

It reinforces how powerful and far-reaching social media terms of service can be in protecting platforms from copyright claims.
It shows the importance of making specific objections and arguments in court—especially during appeals.
It highlights how courts apply procedural rules strictly, even when someone is representing themselves without a lawyer.

Sethunya v. TikTok, 2025 WL 1144776 (10th Cir. April 18, 2025)

December 1, 2024 Artificial Intelligence, Blockchain, Copyright, Cryptocurrency, non-fungible token, Social Media

X can claim trespass to chattel in data scraping case

X Corp. sued Bright Data Ltd. for unauthorized access to X’s servers and the scraping and resale of data from X’s platform. Plaintiff sought the court’s permission to file a second amended complaint after the court dismissed its prior complaint. The court granted plaintiff’s motion in part and denied it in part, allowing some claims to proceed while dismissing others.

Plaintiff alleged that defendant’s scraping activities caused significant harm to its systems. According to plaintiff, defendant’s automated scraping overwhelmed servers, causing system glitches and forcing plaintiff to purchase additional server capacity. Plaintiff further alleged that defendant used deceptive techniques, including fake accounts and rotating IP addresses, to bypass technical barriers and access non-public data. Plaintiff claimed that these actions violated its Terms of Service, interfered with its contracts, and constituted unfair and fraudulent business practices. Plaintiff also introduced new claims under federal and state anti-hacking laws, including the Digital Millennium Copyright Act and the Computer Fraud and Abuse Act.

The court agreed with plaintiff on several points. It allowed claims related to server impairment, including trespass to chattels and breach of contract, to move forward. The court found that plaintiff’s revised complaint provided sufficient details to plausibly allege harm to its servers and unauthorized access to its systems.

However, the court dismissed claims concerning the scraping and resale of data, ruling that they were preempted by the Copyright Act. Plaintiff had argued that it could prevent defendant from copying user-generated or non-copyrightable data through state-law claims. The court disagreed, holding that such claims conflicted with federal copyright policy, which limits protections for factual data and prioritizes public access. Additionally, the court rejected plaintiff’s argument that defendant’s actions constituted “unfair” business practices, finding no evidence of harm to competition.

Finally, the court allowed plaintiff to proceed with its new anti-hacking claims but left the door open for defendant to challenge these allegations later in the case.

Three Reasons Why This Case Matters

Defines Platform Rights: This case clarifies the limits of platform operators’ ability to control user-generated and public data.
Reinforces Copyright Preemption: The decision highlights the importance of federal copyright laws in preventing conflicting state-law claims.
Explores Anti-Hacking Laws: It illustrates how federal and state anti-hacking statutes may be used to address unauthorized access in the digital age.

X Corp. v. Bright Data Ltd., 2024 WL 4894290 (N.D. Cal., Nov. 26, 2024)

Artist’s side hustle lands him in DMCA litigation with autonomous vehicle company

A dispute between a digital artist and his former employer over content rights resulted in a court allowing the employee’s DMCA claim while striking the employee’s state law claims.

For over five years, plaintiff worked for defendant, crafting digital street scenes of San Francisco to train the company’s self-driving vehicles. But plaintiff’s passion for digital art extended beyond his day job. So, in his spare time, using his own equipment, he created intricate urban scenery for video games, which he then sold on the Epic Games marketplace.

When defendant learned of plaintiff’s side hustle, it claimed plaintiff’s project infringed defendant’s copyright rights. It demanded that plaintiff cease all sales of his digital art. Plaintiff refused to comply. He argued that his creations were made on his own time, with his own resources, and did not utilize any proprietary information from defendant.

Defendant considered plaintiff’s refusal as a resignation and terminated his employment. But that did not end the matter. Defendant escalated the situation by sending a takedown notice to Epic Games under the Digital Millennium Copyright Act (DMCA), alleging that plaintiff’s content infringed on defendant’s copyrighted material. This resulted in Epic removing plaintiff’s content from its marketplace.

The lawsuit

Plaintiff sued, claiming that that defendant sent the takedown notice in bad faith, asserting federal claims under the DMCA and state law claims for interference with contractual relations, interference with prospective economic advantage, and violation of California’s Unfair Competition Law. Defendant moved to dismiss the DMCA claim and also moved to strike the state law claims under California’s anti-SLAPP statute, which aims to prevent lawsuits that chill the exercise of free speech.

The court’s decision

The court denied defendant’s motion to dismiss the DMCA claim, allowing plaintiff’s federal claim to proceed. It found that plaintiff had sufficiently alleged that defendant acted in bad faith when it issued the takedown notice, a key requirement under Section 512(f) of the DMCA. But the court granted defendant’s motion to strike the state law claims. It held that the state law claims were preempted by the DMCA and were also barred by California’s litigation privilege, which protects communications made in anticipation of litigation.

Shande v. Zoox, Inc., 2024 WL 2306284 (N.D. Cal., May 21, 2024)