Reaction Videos Have a Fair-Use Problem. The DMCA Has a Process Problem.

Rapid Response / Legal Alert | September 11, 2026

A copyright case ordinarily ends with a judgment.

On YouTube, the consequences can arrive first.

Anime reaction creator Nicholas Light recently said that more than 60 of his videos were temporarily removed after he received more than a dozen copyright strikes arising from takedown requests associated with copyright-enforcement company Remove Your Media. Other anime-focused creators reported similar campaigns. YouTube ultimately removed the strikes against Light, and his channel survived.

That might sound like another installment in the internet’s endless argument over “fair use.”

It is. But it is also something more interesting.

The dispute exposes a structural feature of modern copyright enforcement that matters far beyond anime: a creator can face the economic consequences of losing a copyright case before any court has decided whether infringement occurred at all.

That does not mean the copyright owners are wrong. Some reaction videos use short excerpts to criticize or discuss a work. Others display so much of an episode that viewers can obtain much of the entertainment value without watching the licensed version. Those are very different cases.

But the Digital Millennium Copyright Act does not ask a federal judge to make that distinction before content comes down.

It creates a process.

And in the creator economy, the process can become the punishment.

This distinction gets lost constantly.

YouTube’s automated Content ID system can identify copyrighted material and allow the rightsholder to monetize, track, or sometimes block a video. A Content ID claim ordinarily does not create a copyright strike.

A formal DMCA removal request is different.

When YouTube receives what it regards as a valid copyright takedown request, it removes the identified content and issues a copyright strike. Under YouTube’s current policy, three active copyright strikes within 90 days can subject the channel—and associated channels—to termination.

For a hobbyist, that is unpleasant.

For a creator whose audience, advertising revenue, memberships, sponsorships, archive, and future distribution all live on YouTube, it can be existential.

And here is the first important legal distinction:

Congress did not enact YouTube’s three-strike rule.

Section 512 of the DMCA instead requires a service provider seeking copyright safe-harbor protection to adopt and reasonably implement a policy providing for termination of repeat infringers in appropriate circumstances. YouTube has built its own enforcement architecture around that statutory incentive.

That distinction matters.

Federal law creates strong reasons for platforms to police repeat infringement. But exactly how platforms translate notices into strikes, account restrictions, demonetization, or termination is also a matter of platform governance.

The modern copyright system is therefore not just copyright owner versus creator.

There is a third actor in the room: the platform that controls whether the creator can reach an audience at all.

Section 512 was enacted in 1998, when Congress was trying to solve a very different internet problem.

Online services could not realistically pre-screen everything users uploaded. Copyright owners, meanwhile, needed an efficient mechanism to remove infringing material without filing a federal lawsuit every time someone posted a pirated song or movie.

The compromise was a safe harbor.

Qualifying service providers receive substantial protection from monetary copyright liability for user-uploaded material if they satisfy statutory conditions—including responding expeditiously to compliant infringement notices.

A copyright owner or authorized agent sends a takedown notice identifying the copyrighted work and allegedly infringing material and stating a good-faith belief that the disputed use is not authorized by the owner, its agent, or the law.

The platform removes the material.

The uploader can then send a counter-notification stating, under penalty of perjury, a good-faith belief that the removal resulted from mistake or misidentification. The counter-notice must also disclose identifying information and consent to federal-court jurisdiction.

Once a compliant counter-notice arrives, the service provider generally restores the material after 10 to 14 business days unless the claimant provides notice that it has filed a court action seeking to restrain the allegedly infringing use.

On paper, that looks reasonably symmetrical:

Notice. Removal. Counter-notice. Reinstatement—or litigation.

But economically, it is not always symmetrical at all.

A studio or distributor can send notices as part of a large enforcement operation.

An individual creator must decide whether to submit a legal document, disclose personal information, consent to federal jurisdiction, risk litigation, endure downtime, lose revenue, and potentially expose an entire channel while the dispute is being resolved.

The DMCA moves first and adjudicates later.

That makes sense if the target is obvious piracy.

It becomes much more complicated when the target is a disputed fair use.

Creators sometimes speak about “transformative content” as though adding commentary casts a legal spell over copyrighted footage.

It does not.

Fair use under 17 U.S.C. § 107 requires courts to weigh four considerations: the purpose and character of the new use; the nature of the copyrighted work; how much of the work was taken and how important the portion was; and the effect of the new use on the existing and reasonably expected market for the original.

Reaction content can perform very well under that test.

In Hosseinzadeh v. Klein, a federal court held that Ethan and Hila Klein’s reaction video was fair use where they repeatedly stopped the original video and interspersed clips with substantial criticism and commentary. But the court expressly warned that it was not holding that reaction videos categorically qualify as fair use. The particular video mattered.

That qualification is everything.

Imagine two “reaction” videos.

In the first, the creator plays a 20-second scene, pauses it, analyzes the animation, criticizes a character decision, compares the scene with an earlier episode, and then moves to another excerpt.

In the second, the creator displays nearly an entire episode in sequence while occasionally laughing, nodding, or making short comments.

Calling both “reaction videos” tells us almost nothing legally.

Copyright cases turn on what the secondary work actually does.

Criticism and comment are specifically identified in § 107 as paradigmatic activities that may constitute fair use.

That gives genuine reaction and analysis content a strong starting point.

If the copyrighted clips are being used as the object of criticism—because viewers need to see the scene to understand what the creator is analyzing—the new work may have a meaning and purpose materially different from merely watching the anime itself.

But courts have become increasingly careful about treating “new meaning” as the end of the inquiry.

In Andy Warhol Foundation v. Goldsmith, the Supreme Court emphasized that adding a new message, aesthetic, or expression does not automatically make a commercial use fair. Courts must consider the specific purpose of the challenged use and how closely it resembles the purpose served by the original.

That does not doom reaction videos. Criticism of an anime episode ordinarily serves a meaningfully different purpose from the episode itself.

But it does undermine the simplistic argument that any video becomes fair use merely because a creator put his face beside it and said something.

Commentary matters. What the commentary is doing matters more.

Anime is highly creative expression. That tends to favor the copyright owner under fair-use factor two, although courts often give that factor less weight when a secondary work is genuinely critical or transformative.

The amount used can be far more consequential.

There is no lawful “30-second rule,” “10-percent rule,” or other automatic numerical safe harbor.

A creator may sometimes lawfully use a substantial portion of a copyrighted work when that amount is reasonably necessary to accomplish a legitimate critical purpose. Conversely, taking even a smaller portion can weigh against fair use if the creator takes the expressive heart of the work without adequate justification.

For anime reactions, this creates a continuum.

Short excerpts punctuated by extensive discussion look very different from an uninterrupted episode.

Blurring the screen, shrinking the image, changing opacity, or lowering audio may affect the practical viewing experience, but those techniques do not automatically answer whether protected expression has been copied or whether the resulting work is fair.

And the closer the reaction video gets to allowing the audience to consume the underlying episode, the harder the legal argument becomes.

Remove Your Media’s lawyer told The Verge that some targeted reaction content displayed copyrighted works in their entirety, including content otherwise available behind a paywall, and argued that viewers may watch the reaction instead of a licensed version.

If proven for a particular video, that matters.

Fair use strongly disfavors uses that function as market substitutes for the original.

A reaction creator can reasonably say, “Nobody watches my channel instead of watching the anime; they watch because they already know the show and want my commentary.”

A rightsholder can reasonably respond, “Then why does your audience need to see nearly the entire episode?”

That is the kind of factual dispute courts actually decide.

Analytics could matter. So could the amount of footage shown, availability of the original, whether the reaction preserves the narrative experience, how much original commentary interrupts it, whether viewers need the source material to understand the commentary, and whether the creator offers less-altered or full-length versions through Patreon or another paid service.

That last fact could be particularly significant.

A free YouTube commentary video built around excerpts presents one market question.

A paid product that effectively allows subscribers to watch a copyrighted episode alongside the creator may present another.

Fair use is intensely fact-specific because business models are facts too.

Copyright owners do not receive unlimited discretion simply because fair use is uncertain.

The Ninth Circuit's decision in Lenz v. Universal Music Corp. is especially important here.

Section 512 requires a takedown sender to state that it has a good-faith belief that the challenged use is not authorized by the copyright owner, its agent, or the law.

The Ninth Circuit held that fair use is a use authorized by law for purposes of that provision. A copyright owner therefore must consider fair use before issuing a DMCA takedown notice. Ignoring fair use altogether can create liability under § 512(f), which permits recovery against someone who knowingly materially misrepresents that material is infringing.

That sounds like a powerful safeguard.

In practice, it is narrower than it first appears.

Lenz applies a subjective good-faith standard. If a rightsholder actually considers fair use and forms a genuine belief that the use is infringing, the Ninth Circuit said the court does not simply impose § 512(f) liability because it would have reached the opposite conclusion.

So there is an enormous difference between these propositions:

“The claimant was wrong about fair use.”

and

“The claimant knowingly misrepresented that it had formed a good-faith belief about fair use.”

The first does not necessarily establish the second.

That makes claims of “false copyright strikes” legally more complicated than the phrase suggests.

YouTube's reversal of a strike does not itself establish that a claimant violated § 512(f). Likewise, an enforcement company filing a large number of notices does not, standing alone, establish abuse.

But scale can make the process used to reach the fair-use determination increasingly important.

If enforcement is highly automated, a litigation question practically writes itself: What exactly was reviewed before each statutory certification was made?

This is where the present controversy becomes more interesting than another abstract debate over fair use.

Suppose dozens or hundreds of videos are targeted through substantially identical notices.

For the creators, counsel would want to know whether the videos actually share the relevant characteristics. How much copyrighted footage appeared in each one? How frequently was playback interrupted? What kind of commentary was added? Were full episodes reproduced? Were some videos fundamentally different from others?

For the claimant, the corresponding questions concern process.

Who identified the allegedly infringing works? What authority did the enforcement agent possess? What criteria were applied? Was fair use considered at the individual-video level, by category, through technological rules, or some combination? What records exist showing how the good-faith determination was made?

Those questions matter because copyright litigation frequently turns not on slogans, but on classification.

A batch containing obvious substitutes, close fair-use calls, and obvious criticism should not necessarily receive one legal answer simply because all three categories appear on reaction channels.

That is also why large-scale automated enforcement creates risk for rightsholders themselves. Automation can make copyright enforcement far cheaper. But the more a statutory certification depends on automation, the more important it becomes to demonstrate that the system was designed to identify legally relevant distinctions rather than merely detect copying.

The copyright owner is entitled to protect its work.

It still has to identify infringement.

YouTube told The Verge that it does not decide ownership disputes or independently adjudicate whether challenged videos constitute fair use. When it receives a valid legal removal request, it removes the material; creators who believe the use is lawful can submit counter-notifications.

That posture makes sense.

Fair use routinely requires contextual judgments that judges themselves debate. It would be unrealistic to expect a platform processing enormous volumes of copyright complaints to conduct miniature federal trials.

But that leaves an unavoidable governance problem.

The platform may decline to decide the copyright dispute while simultaneously making decisions about strikes, monetization, account access and termination that determine whether the creator still has a business when the copyright dispute ends.

That distinction deserves more attention.

There are really two questions:

Did the video infringe copyright?

And:

What should happen to the creator's account while that remains genuinely disputed?

The DMCA principally addresses the first problem through notice and counter-notice procedures.

The creator economy has made the second one enormous.

A conventional defendant can decide whether a threatened copyright claim is worth fighting.

A platform creator faces another calculation.

If challenging the claimant creates even a small possibility that the entire channel disappears, the rational business decision may be to stop publishing legally defensible material.

That is a form of what lawyers sometimes call risk asymmetry.

The rightsholder may have thousands of works, an enforcement vendor, institutional counsel, and an established litigation budget.

The creator may have one channel.

The reverse asymmetry can also exist. Rights owners—particularly smaller studios and independent artists—cannot realistically file individual federal lawsuits against every account distributing their work. Without notice-and-takedown procedures, online infringement could become practically unenforceable.

That is why the answer is not simply “make takedowns harder.”

The DMCA is trying to balance two legitimate problems.

The trouble is that the internet it regulates has changed.

The DMCA was built to remove files. The creator economy turned those files into businesses.

It would be easy to reduce the Nicholas Light controversy to a referendum on reaction videos.

That would miss the more important story.

Some reaction videos are probably strong fair uses.

Some are probably not.

A large number live somewhere in between.

That uncertainty is not evidence that copyright law has failed. Fair use is deliberately contextual because criticism, parody, commentary, appropriation and substitution cannot sensibly be governed by a stopwatch.

The problem emerges when a deliberately nuanced doctrine meets an enforcement system built for speed.

A copyright owner needs a mechanism capable of stopping infringement before a pirated work spreads everywhere.

A creator needs some meaningful ability to challenge a questionable removal before the challenge itself destroys the channel.

A platform wants the protections of § 512 without becoming judge and jury in millions of copyright disputes.

Each position is understandable.

Together, they produce a system in which the most important decision can occur before the legal question is ever answered.

And that is what makes the anime reaction fight worth watching.

The immediate controversy may be about Bleach, One Piece, reaction channels and copyright strikes.

The larger question is about something much bigger:

In a creator economy built almost entirely on private platforms, who effectively decides what fair use means—the courts that announce the doctrine, the copyright owners who invoke the DMCA, or the platforms that control whether anyone is still online when the dispute is over?