(Listed in order of most recent to last) 

As we ascertain the parameters of using AI as companies and creators; navigating both the ways in which we can work more efficiently and effectively with AI, but at the same time ensuring that we stay abreast of what the law states in terms of ownership, the landscape is quite literally shifting every day.

As a result, for the first time, we will be providing you with an updated synopsis of the cases that comprise US case law regarding AI and intellectual property ownership. Typically, we never provide actual case citations, but due to the quickly-shifting nature of this new frontier, we have chosen not to gatekeep the information we ourselves use to inform both our clients and ourselves as businesses owners. With that in mind, as always, please keep in mind that this content is for general educational purposes only. If you use AI for any portion of any work you want to own (ie, work you want to be able to protect), please reach out to your attorney to ensure you are complying with current laws. You may contact our firm at any time here.

Authors Guild, et al. v. OpenAI, Inc. Sept. 19, 2023

1. Facts: The Authors Guild including a number of well known authors, including George R.R. Martin, John Grisham, and Elin Hildebrand, brought a complaint against OpenAI and its entities in the U.S. District Court for the Southern District of New York on September 19th for direct, vicarious, and contributory copyright infringement. The complaint alleges that OpenAI has essentially copied Plaintiffs’ works, without permission, into their algorithms and databases which allows their program to output derivative work that “is based on, mimics, summarizes, or paraphrases Plaintiffs’ works.” This, in turn, harms writers and their ability to own their work and make a living from it. The Authors Guild and various writers presented options that OpenAI could have taken to protect authors’ works without violating copyright laws such as paying a licensing fee.  
2. Issue: Can whole books found online through various databases be inputted into AI databases without licensing agreements?
3. Holding: The case has not been decided yet. Plaintiffs are seeking monetary damages for market usurpation and lost opportunity to license their works as well as a permanent injunction to stop OpenAI’s use of their works all together.

Chabon v. OpenAI, Inc. Sept. 8, 2023

1. Facts: A class action lawsuit was filed by a group of writers on September 8th in the Northern District of California against OpenAI and its entity ChatGPT for copyright infringement. The complaint alleges that Open AI incorporates the writers copyrighted works into Open AI’s databases and language learning model to generate “not only summaries, but in-depth analyses of the themes present in Plaintiffs’ copyright works.” The writers state they did not consent to their works being used as “training material” for ChatGPT models and, now, ChatGPT generates a handsome profit for OpenAI’s unauthorized use of their works.
2. Issue: Are copyrighted works allowed to be used as part of training language models and datasets without violating copyright law?
3. Holding: The case has not been decided on yet, but the case has been reassigned to new judge as of October 10th. The claims brought against OpenAI include direct and vicarious copyright infringement along with unfair competition and negligence claims.

J.L., C.B., K.S., et al., v. Google Inc. July 11, 2023

1. Facts: A group of anonymous plaintiffs have brought a class action against Google in the Northern District of California on July 11th for numerous claims including unfair competition, negligence, invasion of privacy, conversion, and various forms of copyright infringement. These claims stem from Google’s alleged routine of using private user data and other web-scraped data to build their AI products like a database called “Bard.” These infringements range from stealing one plaintiff’s entire book for their database, to another’s photographs, and more. Plaintiffs allege their copyrighted works have been incorporated into their database without permission, and Google’s AI system creates illegal derivatives of their works.
2. Issue: Is use of web-scraped data or private user data allowed to be used by companies to build AI systems? Even if the works themselves are copyrighted?
3. Holding: This case has not been decided on yet, but Google filed a Motion to Dismiss against the Plaintiffs on October 16th.

Silverman, et al. v. OpenAI, Inc. July 7, 2023

1. Facts: Famous comedian and actress, Sarah Silverman, along with some other authors have filed a class action against OpenAI on July 7th in the Northern District of California for direct and vicarious copyright infringement, unjust enrichment, unfair competition, and negligence. Plaintiffs and class members are authors of books. The complaint alleges that OpenAI has used the Plaintiffs copyrighted works as “training material” for ChatGPT, and OpenAI “benefit[s] commercially and profit richly” from the use of the copyrighted work. Interestingly enough, the complaint walks through reports from OpenAI regarding the source of their long-form writing sources over the last few models of ChatGPT. OpenAI trained ChatGPT using various sources of unpublished or non copyrighted books from controversial “shadow library” websites. These sources have been disclosed by OpenAI until the most recent paper for the most recent update of ChapGPT which stated no details of their new book sources. Now, Sarah Silverman along with other authors are demanding damages and an injunction related to the allegedly illegal use of their copyrighted works.
2. Issue: Should AI companies be allowed to use “shadow library” type websites to train their programs even if this includes work that is copyrighted?
3. Holding: This case has not been heard yet, but OpenAI has filed a Motion to Dismiss to which the Plaintiffs have replied to. The motion and response have not been heard yet.

Kadrey, et al. v. Meta Platforms, Inc. July 7, 2023

1. Facts: The same plaintiffs of Silverman, et al. v. OpenAI described above filed a similar complaint on July 7 in the same court against Meta, the new name for the Facebook company. The complaint alleges that LLaMa, a large language model used by Meta, is programmed through “copying massive amounts of text and extracting expressive information” from copyrighted works owned by the Plaintiffs. Meta claims that LLaMa uses a training dataset with data that was “publicly available, and compatible with open sourcing.” Open sourcing allows for licensed use of copyrighted work from a open source database, but only if the actual owner gives such license to the database in the first place. Further, Meta also trains their AI using a similar “shadow library” as OpenAI does with purposefully vague sources and explanations of what books are being used for training. This “shadow library” includes Plaintiffs books. 
2. Issue: Should AI companies be allowed to use “shadow library” type websites to train their programs even if this includes work that is copyrighted?
3. Holding: This case has not been heard yet, but Meta has filed a Motion to Dismiss to which the Plaintiffs have replied to. The motion and response have not been heard yet.

Tremblay v. Open AI, Inc. June 28, 2023

1. Facts: Authors, Paul Tremblay and Mona Awad, filed suit against OpenAI for copyright infringement. The authors allege that OpenAI fed their copyrighted material into OpenAI’s system, ChatGPT, without permission. Their books were a part of a larger group of books run through the large language model that powers ChatGPT, all allegedly used without authorization. The Plaintiffs seek damages and an injunction to stop the use of their copyrighted texts. 
2. Issue: Should AI companies be allowed to use copyrighted material to train their AI programs? Or is this a violation of copyright law?
3. Holding: Author, Mona Awad, voluntarily dismissed herself from the case, but the case continues with Paul Tremblay as plaintiff. OpenAI has filed a motion to dismiss, but it has not been heard yet.

Plaintiffs P.M., K.S., et al. v Open AI LP, et al. June 28, 2023

1. Facts: A group of minors sued OpenAI and Microsoft for a violation of numerous laws as well as invasion of privacy, negligence, conversion, and unjust enrichment. OpenAI and Microsoft, collectively, have worked on programs like ChatGPT, Dall-E, and Vall-E. The plaintiffs argue that OpenAI and Microsoft have taken private information from many individuals without consent to input into their AI programs. They argue this goes far beyond authorized use of their personal data that exists online, and their personal information should not be used to train AI programs.
2. Issue: Can private information found through online databases be used by companies to train their AI programs? Does this go beyond the scope of authorized use when people voluntarily put information online?
3. Holding: The Plaintiffs filed a voluntary dismissal on September 15th, so the case will, unfortunately, not be heard. The parties likely settled the case out of court.

Walters v. OpenAI LLC June 5, 2023

1. Facts: Walters brought a case against OpenAI for libel on June 5th in a Georgia state court which OpenAI removed to Federal Court. Essentially, a third party journalist was researching a piece over a lawsuit in which Walters was in no way involved. However, when using ChatGPT to gain information on the case, ChatGPT claimed Walters was the defendant party liable in the case for various claims. Walters is suing OpenAI for making “false and malicious” allegations through ChatGPT that would “injure Walter’s reputation and expos[e] him to public hatred, contempt, and ridicule.”
2. Issue: Can a company be liable for libel when its program gives false information to users about a person? Can a program commit libel? Even if it was just a programming error?
3. Holding: The case has not been heard. OpenAI has filed a motion to dismiss which has also not been heard yet.

Young v. NeoCortext, Inc. Apr 3, 2023

1. Facts: On April 3rd, a class action lawsuit was brought against a company, NeoCortext, Inc., in California federal court for their app, Reface, which allows individuals to swap faces with famous celebrities, athletes, actors, musicians, and other well-known individuals. The app profits off of users by offering various types of subscriptions. The lawsuit is led by Kyland Young, a TV personality, who claims the app is a violation of their right of publicity. Young claims that NeoCortext is required to gain consent before use of the celebrities’ faces, and NeoCortext has not done so. However, NeoCortext argues there is not only no right of publicity claim in this case but also no violation of copyright law. In fact, NeoCortext argues the claims are barred by the First Amendment. 
2. Issue: Can apps freely profit off of the use of well-known individuals likeness?
3. Holding: This case has not been heard yet.

Flora, et al. v. Prisma Labs, Inc. Feb 15, 2023

1. Facts: Prisma Labs, Inc. is a company that develops apps for editing photos and videos. One of their apps, Lensa, allows users to create custom avatars or retouch their own “selfies.” Jack Flora and a group of other individuals have brought suit in California court against the company for violating privacy law. Plaintiffs allege that Prisma Labs does not properly notify users of how their uploaded photos and videos will be used or how the biometric data will be properly destroyed. The app uses an open-source AI program to create the avatars, and many artists and users have complained of their images being used without permission.
2. Issue: Should companies be prohibited from using photos uploaded by users to train their AI programs? Even if the photos were uploaded voluntarily to use their product?
3. Holding: The case is now being moved to arbitration, so it will not be heard in California court anymore.

Getty Images (US), Inc. v Stability AI, Inc. Feb. 3, 2023

1. Facts: Getty Images filed suit against Stability AI in the District Court of Delaware on February 3rd for violations of copyright laws regarding photographs from Getty Images collection. Getty claims that Stability has copied millions of their copyrighted photographs and infringed on their trademarks. Stability AI’s system outputs lower-quality images similar to ones copied from Getty while still using a warped version of the Getty Images watermark. Getty argues this not only violates their intellectual property rights but gives rise to dilution. 
2. Issue: Are AI programs allowed to reproduce copyrighted works that are publicly available online? And include companies’ logos or watermarks on the reproduced works?
3. Holding: The case has not been heard on yet, but some issues with discovery seem to have occurred. Stability AI did file a motion to dismiss which has not been heard on yet.

Andersen, et al. v. Stability AI Ltd., et al. Jan 13, 2023

1. Facts: On January 13, a group of artists brought suit against Stability AI in the Northern District of California for using their artwork without permission to program their AI to create images with striking similarities to the copyrighted art. The plaintiffs are bringing claims for copyright infringement for the “infringing derivative works” but also claims of unfair competition and right of publicity. Stability AI argues that the artwork is publicly available on the internet, and they have not copied the images for distribution.
2. Issue: Can companies take artwork that is publicly available to program their AI, even if the artwork is copyrighted? Do AI created images based on copyrighted work count as derivative works? Can the style of an artist be protected against AI?
3. Holding: The case has not been fully heard yet. Stability AI filed a motion to dismiss, and the judge granted the motion, dismissing all claims except the claim for direct copyright infringement.

Thomson Reuters Enterprise Centre GmbH et al. v. ROSS Intelligence Inc. May 6, 2020

1. Facts: Thomson Reuters, the owner of a legal research platform, brought suit against Ross Intelligence Inc. for copyright infringement and tortious interference with contract. Thomson Reuters research program called Westlaw includes a unique West Key Number System and West Headnotes to assist those conducting legal research. Thomson Reuters alleges that Ross copied the copyrighted Westlaw database, without a license, into their own AI research system to create a competing product. ROSS claimed they only used the Westlaw code to train their program but, in the end, created an entirely new and different code, so there is no infringement of copyrights. ROSS’ motion to dismiss for failure to state a claim was denied.
2. Issue: Can copyrighted programs be used to train AI programs so long as the new code is different? Or is the use of the copyrighted work to train for a new code in and of itself a violation of copyright law?
3. Holding: After denying the motion to dismiss, the case has not been further heard on.