What’s actually slowing this PC down?
Free tools Windows power users keep installed
One-click scans. No signup required.
Pick the symptom - the matching free tool is one click away.
ThatPainter is reader-supported. When you buy through links on our site, we may earn an affiliate commission. Learn More
As an Amazon Associate I earn from qualifying purchases.
As of August 18, 2026, the artists’ lawsuit against Stability AI, Midjourney, DeviantArt, and Runway is still active and has not reached a final ruling or settlement. The case is in discovery, where the parties are fighting over training datasets and technical evidence. The court currently lists a June 1, 2027 trial date, but that date could change. For now, the key question is what the evidence will show—not whether a court has already ruled that AI training infringes copyright.
What case is this?
Andersen et al. v. Stability AI Ltd. et al. is a proposed class action in the U.S. District Court for the Northern District of California, San Francisco Division, case No. 3:23-cv-00201-WHO. Artists filed it on January 13, 2023. Judge William H. Orrick presides over the case; Magistrate Judge Lisa J. Cisneros handles discovery. The named defendants are Stability AI, Midjourney, DeviantArt, and Runway. Plaintiffs include Sarah Andersen, Karla Ortiz, Grzegorz Rutkowski, Gregory Manchess, Gerald Brom, Jingna Zhang, Julia Kaye, and Adam Ellis. The case is proposed as a class action, but a class has not been certified. Artists whose work may have appeared in a training set are not automatically plaintiffs or class members. The court’s case page provides the docket and current case information.
#1 Best Overall
What the artists say the companies did
The plaintiffs’ theories concern several potentially distinct steps: making copies while assembling datasets, using those copies to train or fine-tune models, distributing or commercially exploiting models alleged to contain protected expression, and generating outputs that may reproduce particular works. They also assert artist-name and false-endorsement theories against Midjourney, including allegations about a list of roughly 4,700 artists and the use of artists’ names in product showcase materials.
Those allegations raise different questions that should not be collapsed into a single claim of “style theft.” Whether an image was included in a dataset is different from whether a particular model was trained on it, whether the model retains or can reproduce protected expression, and whether a particular output is substantially similar to a copyrighted work. A prompt that names an artist may raise trademark, false-endorsement, publicity, or unfair-competition issues distinct from copyright. Copyright generally does not give an artist exclusive ownership of a broad style; claims about particular protected expression are a different matter.
What the judge decided—and did not decide
In an August 12, 2024 order, Judge Orrick allowed the principal copyright claims to proceed past the pleading stage and denied Midjourney’s motion to dismiss the artists’ Lanham Act claims involving alleged false endorsement and trade dress. The order dismissed the DMCA claims with prejudice, dismissed unjust-enrichment claims with leave to amend, and dismissed certain breach-of-contract theories against DeviantArt with prejudice. Read the order.
That was a decision about whether the allegations were plausible enough to continue, not a finding of infringement or liability. The court did not rule that AI training is unlawful, that the models contain unlawful copies, that artistic style is protected by copyright, or that the plaintiffs will win. Questions about copying, model operation, outputs, substantial similarity, and fair use remain for the evidentiary stages of the case. The Midjourney artist-list allegations likewise survived a pleading challenge; they are not a judicial finding that naming artists constituted false endorsement.
Why training data is the immediate battleground
The current phase is extended discovery: the process in which each side seeks evidence from the other. A central dispute concerns Midjourney’s training materials. The parties have substantially agreed on identifying and producing final versions of datasets used for its main model versions. Plaintiffs also seek materials described as “Art” and “Artist” datasets, which they believe may relate to training or fine-tuning models and auxiliary systems. Midjourney disputes whether those materials are relevant or within the proper scope of discovery. The discovery filing describes the dispute.
Dataset evidence could help connect a particular artist or work to a particular defendant, dataset, model version, training process, and alleged use. It could also inform the defendants’ likely fair-use arguments. Evidence of deliberate collection or use of artist-specific works may matter to the plaintiffs’ theories; evidence about the purpose and operation of the models may support the companies’ defenses. Neither the appearance of a work in an index nor an artist’s name in a dataset, standing alone, establishes that a defendant downloaded it, used it to train a particular model, or infringed it.
The plaintiffs have also sought information about Midjourney’s “overfit” filter and style-reference technology. Technical evidence about those systems could bear on whether the product was designed to limit memorization or reproduction, or to enable artist-specific imitation. The existence of a discovery request does not prove either proposition; it shows that the parties consider the systems relevant to questions about copying, outputs, intent, and defenses. Confidentiality disputes may also mean that some evidence is not made public.
What the artists would have to establish
At later stages, the plaintiffs will need evidence tailored to the claims and defendants at issue. Depending on the claim, important questions include:
- Ownership and scope: Does a plaintiff own a valid copyright in the work at issue, and is the work registered where registration is required?
- Use by a particular defendant: Was the work actually copied or used by that company, in that dataset, for that model version or process? Public availability alone does not establish permission—or actual use.
- Where copying allegedly occurred: Was the alleged copying in dataset creation, training, model distribution, or a particular output? These are different theories and may require different proof.
- Protected expression and similarity: Does the evidence show protected expression retained or reproduced by a model, or an output substantially similar to a particular work? A resemblance to an artist’s general aesthetic is not, by itself, the same showing.
- Fair use and other defenses: The companies may argue that their uses were fair, that the relevant work was not used, that outputs are not substantially similar, or that a system has substantial non-infringing uses. The outcome will depend on the evidentiary record and applicable law.
- Artist-name claims: For the Midjourney Lanham Act theories, the plaintiffs would need to prove the elements of the claims, including the relevant likelihood of confusion or false-endorsement issues; an artist name appearing in promotional or product material does not decide those questions on its own.
- Class treatment and damages: The plaintiffs must also meet federal class-certification requirements. If the proof is highly individualized by artist, work, model, or output, class treatment could be disputed; damages must be supported as well.
Model version matters throughout. Evidence about Stable Diffusion 1.4, 1.5, XL, or one Midjourney release should not automatically be generalized to a later release. Likewise, a user’s prompt or output may create facts that do not establish company liability.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What happens next
The immediate work is continued discovery and resolution of disputes about datasets, custodians, source code, confidentiality, and expert material. Depositions and expert reports may address training data, model architecture, memorization, outputs, and damages. The parties may then litigate class certification and file summary-judgment motions before trial. This is the likely procedural path, not a guaranteed schedule: extensions and discovery disputes have already affected the calendar.
The docket has recorded continuing activity, including a discovery proceeding on July 23, 2026, a joint discovery letter brief filed by Runway on August 10, and Midjourney sealing-related filings on August 14. A June 5 order extended a deadline concerning Midjourney end-use training data to June 15; a June 17 order further extended a deadline concerning art- and artist-specific datasets to June 23. The docket currently lists June 1, 2027 as the trial date. That is a current setting, not a guarantee. Check the court’s docket page for changes.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →Could it settle instead?
Yes. A settlement could occur before trial, but there is no verified public indication that one is imminent. Possible terms could include payments to named plaintiffs or a broader class, a licensing or compensation arrangement, dataset disclosure or audit obligations, opt-out or removal processes, or changes to training-data sourcing or artist-name features. A settlement might include no admission of liability.
A settlement would resolve some or all of the parties’ dispute on agreed terms, but would not necessarily produce a broad judicial ruling on whether training generative models is fair use. That distinction matters: a negotiated outcome can shape company practices without establishing a precedent on the underlying legal questions.
What Getty’s separate litigation does—and does not—show
Getty Images’ litigation is useful context, not a decision in Andersen. Getty’s U.S. case against Stability AI survived most of Stability’s motion to dismiss in April 2026 and remained in fact discovery, according to Getty’s SEC filing. In the separate UK proceedings, Getty established that its images were used to train Stable Diffusion but did not prevail on its secondary-copyright claim; related proceedings and appeals continue. The UK judgment concerns different proceedings, claims, and evidence. It does not control the U.S. artists’ case.
What to watch for
- Court orders resolving disputes over Midjourney’s datasets, source code, or technical systems.
- Evidence showing whether particular works were used by particular defendants and model versions.
- Depositions and expert reports on memorization, outputs, and damages.
- Class-certification briefing and any ruling on whether the proposed class can proceed.
- Summary-judgment motions, settlement announcements, and changes to the trial schedule.
The consequential near-term development is therefore likely to be what discovery establishes and which claims can be tied to particular works, models, and conduct. Until a merits ruling or settlement, the case remains a live test of disputed copyright and artist-name theories—not a ruling that AI training is categorically lawful or unlawful.
Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallQuick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
