AI
July 7, 2026
By Emily Poler
A couple of months ago, I wrote about a case brought by big movie studios, including Disney, against Nanoble Pte. Ltd. over its Hailou AI app, which allows users to generate short videos in response to prompts. The media empires’ problem with Hailou AI, which Nanoble markets as “a Hollywood studio in your pocket,” is that it can be used to generate videos featuring Star Wars, Marvel, and other well-known, copyrighted characters — videos like the ones Plaintiffs themselves created with Hailou AI to submit as evidence with their filing. And that’s where things suddenly get interesting.
Recently, Nanoble answered the studios’ complaint with an unexpected salvo: a counterclaim alleging the studios breached Nanoble’s terms of service by creating those sample videos and, thus, to the extent Nanoble is found liable for copyright infringement, the studios themselves will have to pay Nanoble’s damages and its attorneys’ fees. In other words, in this seemingly bizarre scenario, if the studios win, they themselves could be liable for any monetary judgments against Nanoble. Talk about hoisted by their own petard! (I think I finally understand what that expression means, even if I still don’t know what a petard is).
How does this make sense? Well, as is generally stated by the terms of service for the apps and websites we all routinely and thoughtlessly sign, by using that service, a user accepts its terms. Here, Hailou AI’s terms of service say users will not use the app in a way that violates applicable law or may expose Nanoble to liability. Notably, Hailou AI’s terms of service provide that if a user violates its terms, that user has to “defend, indemnify, and hold harmless Nanoble against claims, losses, costs, expenses, and fees, including reasonable attorneys’ fees, arising out of or relating to the user’s violation . . . .” In plain English, Nanoble is saying that, because the studios used Hailou AI to create outputs that infringe on their own copyrights, the studios violated Nanoble’s terms of service and are responsible for any monetary judgments against Nanoble.
The studios have responded with a bunch of arguments that amount to saying, in a variety of different ways, that Hailou AI’s terms of service are unenforceable because it would be dumb to allow someone to escape any responsibility for their own actions in this way. They’ve raised a whole host of defenses, including that the counterclaims fail because they violate the California Civil Code which states that “contracts which have for their object, directly or indirectly, to exempt any one from responsibility for … violation of law, whether willful or negligent, are against the policy of the law.” They also assert the counterclaims fail because “Nanoble’s interpretation of its Terms of Use is contrary to the public policies behind Federal Rules of Civil Procedure Rule 11 for pre-lawsuit investigations and the Copyright Act’s policy against copyright infringement and internet service providers interfering with standard technical measures to police infringement online.”
Interestingly, the studios are not claiming that the outputs they created in their Hailou AI videos are not, in fact, infringing, since they own the copyrights to the characters they generated; there are prior decisions holding that a copyright owner can’t infringe on its own copyrights. Going this route could get the studios out of any claim that they breached Nanoble’s terms of service by creating infringing works. However, it would also mean that they couldn’t use the works they generated as a basis for a claim of copyright infringement, which is at the heart of their case. There is also precedent recognizing that the use of copyrighted work in litigation is fair use, but again, doing so could lead a court to hold that since there is no infringement, the studios have no case. See how tricky this is?
One obvious question I have: Couldn’t the studios have found user-generated infringing videos produced using Hailou AI to enter as evidence? Don’t people make Darth Vader videos and post them on YouTube or TikTok? (According to my 14-year-old, yes, duh.)
So where is all this going? Well, Nanoble’s argument is certainly creative. However, I think it’s going to be hard for a court to swallow the idea that a party can basically remove itself from any possibility of liability by a contract. With that said, Nanoble’s arguments significantly increase the studio’s potential exposure, since even if they win, they face the prospect of having to pay Nanoble’s damages and its legal fees. The risk of this may well give the studios incentives to settle and walk away from the whole mess, which seems unfair but may well be better than what would amount to a very 21st-century Pyrrhic victory.
April 29, 2026
By Emily Poler
About a month ago, the Supreme Court in Cox Communications, Inc. v. Sony Music Entertainment reversed a $1 billion verdict against Cox, an Internet service provider. That verdict stemmed from Cox’s failure to prevent subscribers from infringing Sony’s copyrighted works by sharing pirated copies over Cox’s network. The Supreme Court held that a service provider can only be liable for what is called “contributory infringement” by a third-party (in this case, a user) if the service either induced the infringement or sold a service specifically tailored for infringement. In its decision, the Court held Cox was not liable for contributory infringement because it failed to do anything to affirmatively prevent it. Put another way, Cox’s inaction was not enough to incur liability.
One AI company quickly jumped on this decision to argue that it shouldn’t be liable for what users do with its own platform. In Disney Enterprises, Inc. et al v. MiniMax et al, Disney and other large studios are suing China-based MiniMax and its Singaporean owner Nanonoble Pte. Ltd., which offers an app called Hailuo AI. This app allows users to create short videos in response to prompts and is marketed with the tagline, “A Hollywood studio in your pocket.” In theory, this is cool, except that what Hailuo also does is let users make little videos starring Star Wars and Marvel characters, among other well-known properties.
Since Hailuo is AI, naturally the Plaintiffs claim it was created using unauthorized copies of the studios’ works. And, of course, Plaintiffs say any user output that publicly reproduces Darth Vader, Spiderman et al. are unauthorized copies or derivative works of the studios’ copyrighted properties.
MiniMax makes a bunch of arguments as to why the complaint should be dismissed, some of which have enough teeth to maybe lead to dismissal or partial dismissal. For example, MiniMax argues that much of its conduct took place outside of the United States and, therefore, cannot serve as the basis for a claim under U.S. copyright law. In fact, the Plaintiffs spent months after their initial filing unsuccessfully attempting to serve the Defendants in Singapore and China, and the federal judge in the case had threatened to toss the lawsuit in December as a result of these issues.
Other of the Defendants’ arguments are more, shall we say, creative. MiniMax asserts that while the studios complain that Hailuo’s outputs depict characters from their movies and TV shows, the studios can’t point to specific copyright registrations covering those characters. That argument is a bit much because it seems impossible to obtain an image of, say, the Mandalorian or Iron Man, without reproducing a portion of the copyrighted work they appear in. It seems equally hard to claim that a user can make a Shrek short without intruding on the copyright owners’ rights to make derivative works based on their copyrighted materials.
In the wake of the Supreme Court’s Cox decision, MiniMax also argues it shouldn’t be held liable merely because it didn’t prevent Hailuo from outputting infringing content, despite having that ability. (The company did put guardrails in place to prevent Hailuo from creating pornographic or violent content.) According to MiniMax, the failure to include these guardrails is, at most, inaction and not an affirmative act, and within the framework of Cox there’s no contributory liability for mere inaction. However, the line between inaction and action is pretty blurry here. Creating and marketing a platform that is appealing because it lets users make fan films of their favorite characters seems pretty active to me.
On top of that, MiniMax is claiming it’s not responsible for user infringement because it didn’t actively encourage that infringement. Really? Hailuo’s tagline of “A Hollywood studio in your pocket” sounds pretty encouraging to me. MiniMax says the slogan describes “the quality and creative capability” of Hailuo,” but I think it could be interpreted as urging users to create works based on those of Hollywood studios. My view is buttressed by the claim, according to the complaint, that MiniMax allegedly used Spiderman in Hailuo’s marketing materials. This type of question — a fact question — is going to make it hard for the judge to dismiss the complaint.
And that means that despite miniMax’s assertions, there should be plenty of action ahead in this case.
March 19, 2026
By Emily Poler
The last few weeks have been crazy over here (and, obviously, everywhere), so this post will be relatively brief, but interesting nonetheless. It’s about a proposed class action lawsuit initiated by New York Times writer Julia Angwin against Superhuman Platform, the parent company of writing-assistant software Grammarly. While I haven’t had a lot of time to delve into the complaint, the basics are that beginning last August, Grammarly launched its “Expert Review” tool that offered users the ability to revise text according to recommendations from well-known authors like Ms. Angwin, Stephen King, and Neil deGrasse Tyson. Did Grammarly ask permission of these scribes to use their names and writing styles? Of course not!
As you might imagine, Grammarly’s tool is entirely AI-based, having digested the publicly available work of the writers so as to then spit out editing suggestions from what it determines are “relevant experts,” based on the subject matter of uploaded user text. For example, you submit an essay on technology (one of Ms. Angwin’s core subjects) and Expert Review offers improvements it says are “inspired by Julia Angwin.” Pretty amazing stuff, especially considering Grammarly didn’t involve Ms. Angwin in the process.
Even worse than the theft of her style and the use of her name, according to Ms. Angwin, is that the AI advice might (ha! “might”) be crappy. “[A] Grammarly user could become displeased with Ms. Angwin if they…received a negative result after taking that advice (such as a bad grade in school or a negative performance evaluation at work), even though Ms. Angwin had absolutely nothing to do with the advice that she purportedly gave,” the complaint says. Having tested Expert Review herself, Ms. Angwin was appalled by the edits peddled under her name. “Its editing suggestions were so bad that they could destroy my reputation,” wrote Ms. Angwin in a Times opinion piece.
Luckily for Ms. Angwin, even though current laws are, shall we say, in flux regarding the use of AI to digest, learn from, and replicate copyrighted material, her suit rests on far more solid and clearly defined right of publicity laws. These laws, which are on the books in more than two dozen states, bar the commercial use of another’s name or likeness. The authors here have a pretty strong case as Grammarly was using their names without permission. Moreover, it certainly is not crazy for the authors to argue that the use of their names implied that they endorsed Grammarly, which is exactly what the right of publicity is supposed to protect.
How will Grammarly respond? Well, it has already taken down the Expert Review tool “for a redesign,” claiming it had “very little usage.” As for defending itself against the suit, a Superhuman statement says the company “believes the legal claims are without merit.” If you ask me that’s rather optimistic, since it would be quite a stretch to deny that their use of these authors’ names implied endorsement. Grammarly could also try to argue that the authors haven’t been harmed by its use of their names, which might have some legs because reputational damages can be hard to establish. However, such arguments aren’t going to get rid of this case any time soon, nor will the mothballing of the review application which, as Ms. Angwin wrote, “doesn’t make up for the eight months that service was in operation, making money from all of our names without ever seeking our consent.”
January 20, 2026
By Emily Poler
I’ve talked a lot here about the legal implications of AI, whether in copyright infringement lawsuits over its development or problems with how it’s been (mis)used by lawyers. The embarrassment and consequences when an attorney files an AI-drafted brief riddled with hallucinatory errors and false citations? Been there. Copyright infringement cases pending against OpenAI, Meta and other AI companies? Oh yes, we’ve done that. And none of this is ending anytime soon because, no matter how things shake out in the courtroom, one thing is certain: Artificial Intelligence is not going away. If anything, it’s going to become way more pervasive in our business, personal and, yes, legal lives. So with that in mind, let me talk about when, and in what contexts I see AI as a useful tool that can aid legal work — and where I think it’s a bad idea.
Starting with the positives, AI can be great for writing, which doesn’t always come naturally to this human. It can provide a starting point I can then manually edit, which really speeds up getting started on writing tasks that, for whatever reason, I’d just rather avoid. AI is also very useful for repetitive tasks like formatting cumbersome documents like document requests and interrogatories, as well as responses to document requests and interrogatories. (If you’re not a litigator and don’t know what these are, don’t worry. It’s not that exciting.) When it comes to specific AI platforms, in my experience Claude is far better at these routine tasks than Co-Pilot, which could not format things consistently. Hardly surprising, since Co-Pilot is a Microsoft product and despite it now being the second quarter of the 21st century Microsoft still can’t seem to get its most basic product (Word) right, as it still inexplicably changes the formatting of documents without rhyme or reason. But I digress.
How else is AI useful for lawyers? I’ve seen that clients sometimes find AI-generated materials helpful or comforting when they are struggling to comprehend a legal concept. Instead of trying to get me on the phone, they can easily ask ChatGPT relevant questions and get quick answers. Litigation can be quite anxiety-ridden for a client, and if gaining a better understanding of what’s happening puts their minds at ease, fantastic. Of course, we have to keep the big caveat in mind: As everyone should know by now, AI-generated information is NOT always accurate.
Speaking of which, AI use is obviously a real problem when, for example, a lawyer’s time (and billing) is devoted to reviewing bogus legal citations that AI has magically created or when AI produces a case or a statute that says something that seems pertinent, but is provided without the full context and upon further review turns out to be irrelevant. Also, at least in my experience, none of the AI platforms are particularly good at telling when someone is lying or heavily shading the truth. If an adversary is blatantly presenting untrue “facts,” AI platforms — which work by analyzing what words go together — can’t necessarily tell the difference between truth and fiction. It also can’t account for human behavior which, you might have noticed, is sometimes weird and unpredictable.
Time and time again, we see explicit and often embarrassing examples of why AI should not and cannot be trusted by lawyers for research. I’ve written about several cases where lawyers were humiliated and punished by judges for presenting briefs filled with AI-generated nonsense, sometimes digging themselves deeper holes with ridiculous excuses and justifications (here’s an excellent example). And yet, despite this, the use of AI to conduct legal analysis is becoming increasingly prevalent for those who work both inside and outside the legal field. It saves time, it saves money, it makes things easy, and as we know all too well, humans are always eager to overlook errors for the sake of convenience. But I will not get sucked into its wanton and irresponsible use. I might use it for routine and mechanical tasks, but whenever a situation requires critical thinking or multiple logical steps, I rely on hard work and human analysis and forgo the assisting “skills” of generative AI.
One final note: Trachtman & Poler Law is a small firm. I am aware that BigLaw firms have developed their own AI platforms, and the data in these private AI platforms is, well, private. We don’t have that. There may be a time and a place where this is something we explore, but we’re not there yet.
October 14, 2025
By Emily Poler
I previously wrote about the US Copyright Office’s policy on works created with AI and the decision in Thaler v. Perlmutter, which denied copyright registration for a work listing an AI platform as its sole author. In that post, I predicted we’ll soon see litigation over which elements of work created with AI can be copyrighted.
While I’m pretty sure those suits will start to pop up, right now I want to talk about another case where the Copyright Office decided that a work created with AI was ineligible for copyright protection. This case, Allen v. Perlmutter, also raises some of the issues I noted in another recent post where I suggested it might be time to reconsider some of the policies underlying US copyright law in light of how much has changed since the US Constitution and the first copyright law were created in the 18th Century.
The story: Jason Allen created an image titled Théâtre D’opéra Spatial using Midjourney AI and entered it in the 2022 Colorado State Fair’s annual fine art competition, where it won a prize. The US Copyright Office, however, was less impressed and denied his application for copyright protection, finding that it was created by Midjourney. Allen then filed suit challenging that decision. (Before diving in, two notes. One, H/T to Paul LiCalsi for pointing this case out to me. Two, in case you’re wondering, Shira Perlmutter, the defendant in both Thaler and Allen was, until recently, the Director of the US Copyright Office).
Some background. To be eligible for a copyright, a work must be “original” and have an “author.” Of course, the law has long recognized that humans create copyrightable materials using machines all the time. In 1863’s Burrow-Giles Lithographic Co. v. Sarony, the Supreme Court found Napoleon Sarony’s photograph of Oscar Wilde was eligible for copyright protection, rejecting Plaintiff’s argument that photography is a mechanical process devoid of human authorship. The Court ruled that Sarony’s numerous creative choices in composing the photo meant he was the author of the work and, therefore, should be treated as such under the Copyright Act. Since then, courts, including the Supreme Court, have repeatedly held that only a minimal degree of creativity is required for something to be copyrighted.
In this present case, Allen created his artwork by inputting many, many text prompts (over 600!!) into Midjourney to get the result he wanted out of the AI. Also, once he finished creating that initial image, he tweaked and upscaled it using additional software like Adobe Photoshop. The Copyright Office, nonetheless, denied registration for this work, finding that it lacked the “traditional elements of authorship” because Allen “did not paint, sketch, color, or otherwise fix…” any portion of the image.
However, as Allen’s attorney points out in his lawsuit, there is no legal definition of the “traditional elements of authorship” and, what’s more, creativity, not the actual labor of producing a work, is the hallmark of authorship under the Copyright Act.
What to make of this case? Well, for starters, I’m curious to see the Copyright Office’s response regarding its narrow and archaic “traditional elements of authorship.” I imagine it’s going to be hard, if not impossible, to claim those can’t include use of a machine because, well, most everything that is obviously eligible for copyright protection in the 21st Century (music, movies, photography, etc.) uses hardware and software. Also, I wonder the extent to which some of the issues in this case reflect a basic uncertainty about how to characterize and appraise the skills (conceiving and refining detailed prompts) Allen employed to get Midjourney to create the work, compared to what we traditionally think of as visual art skills (painting and drawing). And, elaborating on that last point, how do we define creativity in light of all of the crude AI slop out there? (One example: check out the chair in this clip when the reporter retakes her seat.) Do we need to make some big decisions about what qualifies as helping “to promote the Progress of Science and useful Arts” (the purpose of the Copyright Act) by taking into account that some created work is good, borne of inspiration, purpose and ever-evolving skills, while a lot of stuff that gets made is just plain lazy, bad and crudely functional? Tough calls lie ahead.
