The Midjourney lawsuit is one of the most significant copyright cases in AI history, and in 2026, it is still very much alive. Artists filed a class action against Midjourney, Stability AI, and other AI companies, claiming these platforms scraped billions of copyrighted images to train their systems without permission, payment, or consent.
This case directly affects millions of professional artists, photographers, and illustrators around the world. If your work was published online before 2023, there is a real chance it ended up in an AI training dataset.
In this article, you will find the full case history, a 2026 status update, a breakdown of the legal arguments on both sides, who may qualify, and what any potential settlement could look like.
One number to keep in mind: the LAION-5B dataset used to train these AI systems contains roughly 5.85 billion image-text pairs scraped from the web. That is an almost incomprehensible amount of creative work taken without asking.
What Is the Midjourney Lawsuit?
The Midjourney lawsuit is a federal copyright class action claiming that Midjourney and other AI art companies used billions of copyrighted images without permission to train their artificial intelligence systems.
The formal case is called Andersen v. Stability AI Ltd., filed in the U.S. District Court for the Northern District of California. Case number is 3:23-cv-00201. It targets Midjourney Inc., Stability AI, and DeviantArt as co-defendants.
The core claim is simple: these companies built their products using other people’s creative work. They never asked. They never paid. They never even acknowledged the artists whose work fed their systems.
| Case Detail | Information |
|---|---|
| Case Name | Andersen v. Stability AI Ltd. |
| Case Number | 3:23-cv-00201 |
| Court | U.S. District Court, Northern District of California |
| Filed | January 2023 |
| Defendants | Midjourney Inc., Stability AI Ltd., DeviantArt |
| Lead Attorney | Joseph Saveri Law Firm |
| Legal Claims | Copyright infringement, DMCA violations |
The lawsuit was first filed in January 2023. An amended complaint followed later that year, sharpening the legal claims and adding more specifics about how the defendants allegedly stripped copyright information from scraped images.
Midjourney Lawsuit 2026 Update: Where Does the Case Stand?
As of 2026, the Midjourney lawsuit is in an advanced litigation phase, with key rulings having shaped which claims survived and which were dismissed.
The case went through major procedural hurdles in 2023 and 2024. The court dismissed several claims from the original complaint, but the plaintiffs filed an amended complaint that reinstated and strengthened the core copyright infringement and DMCA arguments.
By 2025, the case had moved into discovery, meaning both sides were required to exchange evidence, internal communications, and data about how the AI training process worked. This is often where class action cases gain or lose momentum.
Key 2026 milestones to watch:
- Class certification hearing (whether the court agrees this is a proper class action)
- Expert witness depositions on AI training methodology
- Potential mediation or settlement discussions
- Trial date, if no settlement is reached
The outcome of the class certification hearing is especially important. If the court certifies a class, tens of thousands of artists could become part of the lawsuit automatically.
What Is the Midjourney Copyright Lawsuit Actually Claiming?
The Midjourney copyright lawsuit claims that the defendants violated U.S. copyright law by copying, storing, and using protected creative works without authorization.
This is not a vague or theoretical claim. The plaintiffs argue that Midjourney’s AI model was literally trained on copies of their artwork. Every time the system processed an image during training, that was an unauthorized copy of a protected work.
The amended complaint focuses on three main legal theories:
| Legal Theory | What It Claims |
|---|---|
| Direct Copyright Infringement | AI training involved copying protected images without a license |
| DMCA Section 1202 Violation | Copyright management info (artist names, metadata) was stripped from images |
| Vicarious/Contributory Infringement | Defendants profited from and enabled the infringement |
Think of it like a musician sampling thousands of songs without clearing the rights, then selling the resulting album for millions. The mechanics are different, but the legal principle the plaintiffs are arguing is similar.
The plaintiffs also argue that AI-generated outputs are not truly “transformative” because they are statistically derived from the original training images.
Who Filed the Midjourney Class Action?
Three artists originally filed the Midjourney class action: Sarah Andersen, Kelly McKernan, and Karla Ortiz.
All three are professional artists with established careers. Sarah Andersen is a widely known cartoonist and author. Kelly McKernan is a fine art painter whose distinctive fantasy style made their work particularly recognizable in AI output tests. Karla Ortiz is a concept artist and illustrator who has worked in film and entertainment.
They were represented by the Joseph Saveri Law Firm and attorney Matthew Butterick, who has been one of the most vocal legal voices on AI copyright issues.
Why these three artists?
- Their styles were demonstrably reproducible by AI art tools
- They had clear copyright registrations for their work
- They represented different categories of visual art, strengthening the class argument
The case is styled as a class action, meaning these three artists are the named plaintiffs representing potentially hundreds of thousands of artists in the same position.
Midjourney DMCA Violation Claims Explained
The DMCA violation claims in this lawsuit may actually be stronger than the core copyright claims, and that matters for how the case resolves.
Under Section 1202 of the DMCA, it is illegal to remove or alter “copyright management information,” or CMI. This includes the artist’s name, the work’s title, copyright notices, and metadata embedded in image files.
The plaintiffs allege that when AI companies scraped images using the LAION dataset, they stripped this CMI from the files. The images were processed and stored without the attached copyright information identifying the rightful owner.
Why Section 1202 matters:
- It carries statutory damages of $2,500 to $25,000 per violation
- The violation does not require proving the defendant knew they were infringing
- Each scraped image could represent a separate violation
| DMCA 1202 Detail | Amount |
|---|---|
| Minimum statutory damages per violation | $2,500 |
| Maximum statutory damages per violation | $25,000 |
| Number of images in LAION-5B dataset | ~5.85 billion |
| Potential exposure if violations proven | Theoretically enormous |
Those numbers are staggering when multiplied across billions of images. This is one reason the defendants have fought hard to have the DMCA claims dismissed.
Key Takeaway: The DMCA violation claims could produce massive damages even if the core copyright infringement arguments face challenges, because each scraped image may count as a separate violation carrying up to $25,000 in statutory penalties.
Midjourney’s Fair Use Defense: Does It Hold Up?
Midjourney’s primary defense is fair use, arguing that using images to train an AI is a transformative purpose that does not require licensing or payment.
Fair use is a legal doctrine that allows limited use of copyrighted material without permission under certain circumstances. Courts weigh four factors:
| Fair Use Factor | Midjourney’s Argument | Plaintiffs’ Counter |
|---|---|---|
| Purpose and character of use | Training is transformative, not expressive | Training copies entire works for commercial gain |
| Nature of the copyrighted work | Facts and ideas cannot be protected | Creative artistic works get strong protection |
| Amount of work used | Only used for pattern recognition | Entire images were copied and stored |
| Effect on the market | AI doesn’t replace original art | AI competes directly with artists for commissions |
Courts have not yet issued a definitive ruling on whether AI training constitutes fair use. That question is the central legal battleground in 2026.
The Google v. Oracle case (2021) established that certain transformative uses of software code could qualify as fair use. Midjourney and its co-defendants lean on that precedent. But the plaintiffs argue that creative visual art is categorically different from functional software code.
Legal scholars are split. This uncertainty is one reason the case has dragged on without a quick dismissal.
How Is the Stability AI Lawsuit Connected to Midjourney?
The Stability AI lawsuit and the Midjourney lawsuit are part of the same case, making this a multi-defendant action targeting the broader generative AI industry.
Stability AI is the company behind Stable Diffusion, one of the most widely used open-source AI image generators. Midjourney itself runs on a model closely related to Stable Diffusion technology. The two companies are linked not just legally but technically.
Key connections:
- Both Midjourney and Stability AI used the LAION dataset to train their models
- LAION scraped images from the public internet at massive scale
- DeviantArt, also named in the suit, operates a platform where many of the original artists published their work
The plaintiffs argue this shared foundation means all defendants participated in the same underlying infringement. A ruling against Stability AI on training data use would likely apply to Midjourney as well.
Stability AI has faced separate financial pressures, including executive departures and reports of funding difficulties. Whether the company can sustain prolonged litigation through 2026 and beyond is an open question.
Midjourney Lawsuit Timeline: From Filing to 2026
The Midjourney lawsuit has moved through several distinct phases since it was first filed in January 2023.
| Date | Event |
|---|---|
| January 2023 | Original complaint filed, Andersen v. Stability AI |
| April 2023 | Defendants file motions to dismiss |
| October 2023 | Court partially grants motions to dismiss, allows amended complaint |
| November 2023 | Amended complaint filed with stronger legal theories |
| 2024 | Discovery phase begins; both sides gather evidence |
| Early 2025 | Expert witnesses retained; class certification briefing begins |
| Mid-2025 | Class certification hearing held |
| Late 2025 | Court issues ruling on class certification |
| 2026 | Case moves toward trial or settlement negotiations |
The partial dismissal in October 2023 was a significant moment. The court found that some of the original claims were not specific enough. But the court gave the plaintiffs a chance to fix that, which they did with the amended complaint.
The amended complaint is more detailed and more legally precise. Legal observers generally view it as stronger than the original filing.
Key Takeaway: The Midjourney lawsuit did not die after the initial partial dismissal. The amended complaint revived the core claims, and by 2026 the case is in its most consequential phase yet.
Who Qualifies for the Midjourney Lawsuit?
Artists who may qualify for the Midjourney lawsuit are those whose copyrighted visual works were published online before the defendants’ training data was assembled, primarily before 2023.
This is not a simple claim to make. You need to meet several criteria to have a realistic argument:
General eligibility indicators:
- You are a visual artist, illustrator, photographer, or graphic designer
- You published original work online before 2023
- Your work is protected by copyright (automatically at creation in the U.S.)
- Your artistic style or specific works appear reproducible by Midjourney or Stable Diffusion
- Your work may have appeared on platforms indexed by LAION, including DeviantArt, ArtStation, Pinterest, or general web search
| Eligibility Factor | Details |
|---|---|
| Work type | Visual art, illustration, photography, digital art |
| Publication window | Published online before 2023 |
| Copyright status | Must be original work (copyright exists automatically) |
| Registration benefit | Registered works may qualify for higher damages |
| U.S. requirement | Case is U.S. federal court; international artists may have separate claims |
Copyright registration is not required to qualify, but it matters for damages. Registered works can claim statutory damages. Unregistered works are limited to actual damages, which are harder to prove.
Artists outside the U.S. may face additional hurdles. The lawsuit is filed in U.S. federal court and U.S. copyright law governs the claims.
Can Artists Sue Midjourney Individually?
Artists can pursue individual copyright infringement claims against Midjourney outside of the class action, but the class action provides a far more practical path for most people.
Filing an individual federal copyright lawsuit is expensive. Legal fees can reach tens of thousands of dollars before a case gets anywhere near trial. For an individual artist, that cost is almost never justified by the potential recovery.
The class action solves this problem by spreading costs across a large group of plaintiffs. If you are part of the certified class, you do not need to hire your own attorney or file your own case.
Options for artists in 2026:
| Option | Cost | Practical for Most Artists? |
|---|---|---|
| Join the class action | No upfront cost | Yes |
| File individual federal lawsuit | Very high | Only for major commercial artists |
| File DMCA takedown notice | Free | Addresses outputs, not training |
| Opt out of class to pursue own case | Personal legal costs | Only if damages are very large |
The class action is your most accessible option if your work qualifies. Monitoring the case for a class certification approval in 2026 is the first step.
Midjourney Lawsuit Settlement Amount: What Are the Numbers?
No official Midjourney settlement has been announced as of 2026, but legal analysts have discussed a range of possible outcomes based on comparable cases and statutory damages exposure.
This is where things get complicated. The potential damages in this case are theoretically enormous. If the DMCA Section 1202 claims survive and the class is certified, the exposure for defendants could run into billions of dollars.
In reality, most class action settlements do not result in maximum statutory damages. Both sides negotiate to a number that the defendant can pay and the plaintiffs’ attorneys can defend to class members.
Illustrative settlement scenarios (not confirmed figures):
| Scenario | Estimated Per-Artist Payout | Notes |
|---|---|---|
| Small settlement, large class | $50 to $500 | Most comparable tech class actions land here |
| Moderate settlement, documented harm | $500 to $5,000 | Artists with registered works and clear evidence |
| Large individual claim, registered work | $5,000+ | Rare; reserved for artists with major commercial losses |
| Statutory damages at trial | Up to $150,000 per work | Willful infringement finding required |
The actual outcome depends heavily on whether the class gets certified, whether DMCA claims survive, and whether both sides prefer settlement over a costly trial.
Key Takeaway: The per-artist payout in a Midjourney settlement will likely range from a few hundred to a few thousand dollars for most class members, with higher amounts possible for artists with registered works and documented economic harm.
Midjourney Lawsuit Payout for Artists: How Would It Work?
If a settlement is reached in the Midjourney lawsuit, the payout process would follow a standard class action distribution model.
Here is how that typically works:
Step 1: Settlement is announced. The parties reach an agreement and submit it to the court for approval.
Step 2: Class members are notified. A notice is sent by mail or email to known class members. It also appears in legal notices and press coverage.
Step 3: You file a claim. Class members submit a claim form with proof of eligibility, typically documentation of published works and copyright ownership.
Step 4: Opt-out deadline. If you want to pursue your own lawsuit instead, you must opt out before the deadline. Missing this date locks you into the class settlement.
Step 5: Court approves the settlement. A fairness hearing is held. The court decides if the settlement is fair, reasonable, and adequate.
Step 6: Payments go out. Payments are distributed, often months after final approval.
| Payout Process Step | What You Need to Do |
|---|---|
| Watch for settlement announcement | Monitor court records and legal news |
| File a timely claim | Submit required documentation |
| Provide proof of published work | Screenshots, platform links, registration certificates |
| Meet the opt-out deadline | Decide before the deadline if you want separate action |
Start gathering documentation of your published works now. Waiting until a settlement is announced creates a scramble.
Midjourney Lawsuit Outcome and Status in 2026
The 2026 status of the Midjourney lawsuit shows a case that has survived early dismissal attempts and is now heading toward its defining moment.
The biggest open question in 2026 is class certification. If the court certifies the class, the lawsuit transforms from a three-plaintiff case into a mass action representing potentially hundreds of thousands of artists worldwide. That changes the settlement math dramatically for both sides.
If the court denies class certification, the named plaintiffs can still pursue their individual claims. The case does not disappear, but it becomes much less powerful as a vehicle for industry-wide change.
Current case status indicators:
- Core copyright infringement claims are active
- DMCA Section 1202 claims survived amended complaint review
- Discovery is ongoing or recently completed
- Class certification ruling is a 2026 watershed moment
Legal observers across the IP world are watching this case. Its outcome will set precedent for every AI company that trained on internet-scraped data, which is essentially all of them.
AI Training Data Copyright: The Legal Issue at the Heart of the Case
The central legal question in the Midjourney lawsuit is whether using copyrighted images to train an AI system constitutes copyright infringement under U.S. law.
This question has never been definitively answered by a federal appellate court. That is what makes this case so important. It is not just about Midjourney. It is about the entire legal framework governing how AI systems are built.
Two competing views:
| Position | Argument |
|---|---|
| Artists’ position | Copying an image to train AI is infringement, same as any other unauthorized copy |
| AI companies’ position | Training is transformative; AI learns concepts, not images; it’s like reading books to learn writing |
The U.S. Copyright Office has weighed in, noting that AI-generated content generally does not qualify for copyright protection. But the office has been more cautious about the training data question.
Congress has held hearings but has not passed legislation addressing AI training data as of 2026. That means the courts are making the law as they go.
The LAION Dataset and Its Role in the Midjourney Lawsuit
The LAION dataset is the engine underneath this entire lawsuit, and understanding it explains why so many artists are affected.
LAION, which stands for Large-scale Artificial Intelligence Open Network, is a German nonprofit that assembled massive datasets of image-text pairs by crawling the public web. The two key datasets are LAION-400M (400 million pairs) and LAION-5B (5.85 billion pairs).
These datasets were used to train Stable Diffusion, which in turn powered or influenced Midjourney’s early models. The images in LAION came from everywhere: personal websites, social media, art platforms, stock photo sites, and online portfolios.
LAION dataset quick facts:
- LAION-5B contains approximately 5.85 billion image-text pairs
- Images were scraped from publicly accessible URLs
- No permission was sought from copyright owners
- Artists have found their specific works in LAION using tools like HaveIBeenTrained.com
| Dataset | Size | Source | Permission Sought? |
|---|---|---|---|
| LAION-400M | 400 million pairs | Web crawl | No |
| LAION-5B | 5.85 billion pairs | Web crawl | No |
The plaintiffs argue that LAION’s scraping, combined with Midjourney and Stability AI’s use of those datasets, created an unbroken chain of infringement from the original artwork to the commercial AI product.
Key Takeaway: The LAION dataset connects every artist whose work appeared anywhere online to the Midjourney lawsuit, because their images may have been included in AI training data without consent or compensation.
What the Midjourney Lawsuit Means for AI Art in 2026
The Midjourney lawsuit in 2026 is reshaping how the entire AI art industry operates, regardless of how the case ultimately resolves.
Major AI companies are already changing their behavior. Some have introduced opt-out mechanisms allowing artists to request removal of their work from future training datasets. Others have begun exploring licensing deals with stock photo agencies and creative platforms.
These changes are happening not because the law requires it yet, but because the litigation risk is forcing companies to act.
Industry shifts triggered by the lawsuit:
- Adobe Firefly was built on licensed images, explicitly to avoid this legal exposure
- Getty Images filed its own separate lawsuit against Stability AI over image scraping
- The U.S. Copyright Office launched a formal study on AI and copyright in 2023, with findings that continue to influence policy in 2026
- Several countries, including those in the EU, have introduced or proposed AI-specific copyright legislation
For artists, the message is clear. Document your work. Register your copyrights. Use available tools to check whether your images appear in training datasets.
For the AI industry, the message is equally clear. The era of scraping the internet for free training data without legal consequence may be ending.
The Midjourney lawsuit is not just a court case. It is a line in the sand between an industry that built itself on borrowed creative work and the people who created that work.
Frequently Asked Questions
What is the Midjourney lawsuit about?
The Midjourney lawsuit is a federal copyright class action claiming that Midjourney and other AI companies used billions of copyrighted artworks to train their AI systems without permission.
The case is formally called Andersen v. Stability AI, filed in the Northern District of California in January 2023.
Plaintiffs argue this constitutes both direct copyright infringement and violation of the DMCA’s prohibition on stripping copyright management information from creative works.
Who qualifies to join the Midjourney class action lawsuit?
Visual artists, illustrators, photographers, and digital creators who published original work online before 2023 may qualify.
Your work does not need to be registered to be part of the class, but registered works can claim higher statutory damages.
The class has not been officially certified as of early 2026, so eligibility criteria will be formally defined by the court’s certification ruling.
How much could artists receive from a Midjourney settlement?
No settlement has been announced as of 2026, but comparable tech class actions typically produce payouts ranging from $50 to $500 per class member for general claims.
Artists with registered copyrights and provable economic harm could receive significantly more.
Statutory damages at trial could reach up to $150,000 per work for willful infringement, though that would require winning at trial rather than settling.
What is the current status of the Midjourney lawsuit in 2026?
The Midjourney lawsuit is in an advanced stage in 2026, with discovery largely complete and class certification proceedings underway.
The court’s ruling on class certification is the most critical near-term milestone, determining whether the case can proceed as a mass action.
If the class is certified, the case gains enormous leverage toward a settlement or trial that could affect artists across the entire AI-generated art industry.
Did Midjourney win the fair use argument in court?
Midjourney has not won a definitive fair use ruling as of 2026.
Courts have not issued a final decision on whether AI training constitutes fair use under U.S. copyright law.
The fair use question remains the central unresolved legal issue, and a ruling either way will set major precedent for every AI company in the industry.
Where This Case Goes From Here
The Midjourney lawsuit is not a minor tech dispute. It is a fundamental question about who owns creative work and whether corporations can use that work freely to build billion-dollar products.
If you are an artist, start documenting your published work today. Keep records of when your work was published, on which platforms, and whether it carries copyright notices.
When a class action settlement notice arrives, and it may arrive in 2026 or shortly after, having that documentation ready will make the claims process faster and your recovery stronger.
The case continues. The stakes are real. And for the first time in the AI era, artists have a serious legal argument in a serious federal court.
