1. The Intercept Media v. OpenAI
Background: The Intercept Media, a news publisher represented by the same firm that represents the plaintiffs in the Raw Story Media litigation below, filed suit against OpenAI and Microsoft on February 28th, 2024 in the Southern District of New York, the same day Raw Story Media commenced their suit. Like the Raw Story allegations, The Intercept alleged that their articles were used to train ChatGPT and brought claims for the removal of the copyright management information (“CMI”) from the articles.
Current Status: MAJOR UPDATE! The Court grants Microsoft’s motion to dismiss with prejudice. The Court also partially grants OpenAI’s motion to dismiss. CMI removal action remains. No rationale given—yet! In a short order with no explanation or reasoning, the Court granted Microsoft’s motion to dismiss with prejudice, and partially granted OpenAI’s motion to dismiss (with prejudice for the 17 U.S.C. § 1202(b)(3) claim but allowing the 17 U.S.C. § 1202(b)(1) to proceed past the motion-to-dismiss stage). The Court stated the reasons for dismissal will issue “in due course.” We will all have to wait to see what precedent, if any, the Court will set with this ruling. Either way, the survival of the CMI-removal claim is a major shift from the dismissal in Raw Story Media. We believed that the court in Raw Story Media got it wrong, so that claim’s survival here is a welcome update.
2. Raw Story Media & Alternet v. OpenAI
Background: News publishers Raw Story Media and Alternet filed suit against OpenAI and Microsoft on February 28th, 2024 in the Southern District of New York, claiming their articles were used to train the LLM that powers OpenAI’s ChatGPT. Rather than claiming copyright infringement, the plaintiffs alleged one cause of action for violating the DMCA (which is a separate provision of the Copyright Act related to Internet content). The plaintiffs claimed that OpenAI removed the CMI from their articles, which they argue is a violation of the DMCA.
Current Status: MAJOR UPDATE. Plaintiffs ask for another shot. As previously reported, the Court dismissed this case for lack of standing, finding that the plaintiffs failed to allege a cognizable damages theory. On November 21st, the plaintiffs requested leave to file an updated complaint. They claim that in this First Amended Complaint, they will allege that the removal of CMI from copyrighted articles constitutes copyright infringement, and not just an injury analogous to copyright infringement, which was rejected by the Court. They also plan to add claims for unjust enrichment. Stay tuned to see whether the Court gives them another chance. When the Intercept media Court provides the reasoning for the survival of the CMI-removal claim in that case, we may see a development on this case as well.
3. The New York Times v. Microsoft & OpenAI
Background: This is the most closely watched litigation involving copyright owners and generative AI tech companies.
On December 27, 2023, The New York Times sued Microsoft and OpenAI in the U.S. District Court for the Southern District of New York for copyright infringement and other related claims. The Times alleges that the companies used “millions” of its copyrighted articles to train their AI models without its consent. The Times claims this has resulted in economic harm by pulling users away from their paywalled content and impacting advertising revenue. The complaint alleges several causes of action, including copyright infringement, unfair competition, and trademark dilution. In its pleadings, The Times asserts that Microsoft and OpenAI are building a “market substitute” for its news and further that their AI generates “hallucinations” based on The Times’ articles also substantially damage its reputation and brand. The Times seeks “billions of dollars of statutory and actual damages.” Microsoft and OpenAI assert the defense of “fair use” - i.e., no license, payment or consent is needed.
On September 13, 2024, the Court granted a motion to consolidate the case with one brought by the Daily News and other publications. The judge assigned is Judge Sidney Stein.
Current Status: MAJOR UPDATE. The Court says NYT’s use of generative AI is not relevant: OpenAI spoliation accident. On November 22nd, the Court issued a major opinion regarding the scope of discovery. Back in September, OpenAI asked the Court to compel plaintiff The New York Times to produce information related to The Times’s use of generative AI, its creation of its own AI products, and its general position on generative AI. OpenAI claimed the discovery was relevant to its fair use defense. Magistrate Judge Ona T. Wang ruled for The Times, denying OpenAI’s motion to compel. The Judge stated that this case is not about the benefits of generative AI or the plaintiffs’ use AI products outside the scope of this litigation. This order settles one issue among several, in what has been a busy and difficult discovery process.
In other discovery news, the plaintiffs notified the Court on November 20th that a significant amount of the data they had collected regarding OpenAI’s training datasets had been inadvertently erased by OpenAI engineers. OpenAI provided the plaintiffs two “dedicated virtual machines” to assist with their searching OpenAI’s training datasets for copyrighted works. Apparently by mistake, OpenAI erased all the data on one of the machines, rendering weeks worth of the plaintiffs’ work useless. This follows several other discovery issues raised by both parties over the past week. All parties will meet on December 3rd, 2024 for an in-person status conference, so we expect new discovery orders to follow.
4. The Center for Investigative Reporting v. OpenAI
Background: The Center for Investigative Reporting, which produces Mother Jones and Reveal, sued Microsoft and OpenAI for essentially the same claims made in The New York Times case above.
Current Status: Motion to consolidate recently granted! Microsoft and OpenAI asked the court to consolidate this case with NY Times v. Microsoft & OpenAI, and — and, as indicated above, the Magistrate granted the motion to consolidate on October 30th (refer to the activity discussed above).
5. Kadrey et al. v. Meta
Background: This case is similar to the “In re OpenAI ChatGPT Litigation” class action below against OpenAI. In this case, Kadrey, comedian Silverman, and others sued Mark Zuckerberg’s Meta on July 7, 2023 in the U.S. District Court (Northern District of California) for mass infringement - i.e., unlicensed “training” of their generative AI model on millions of copyrighted works, including their own. Not surprisingly, Meta’s defense is “fair use.” The judge assigned is Judge Vince Chhabria.
Much like the class action “In Re OpenAI ChatGPT Litigation” above, and for similar reasons, in November 2023, the Court dismissed the bulk of plaintiffs’ claims against Meta. But much like in the OpenAI case, the Court gave the plaintiffs a chance to amend their complaint to add a more direct link to actual harm (and they filed their amended complaint in December 2023).
Current Status: The Magistrate Judge ruled that plaintiff’s discovery requests came too late. Plaintiffs appeal. On November 15th, Magistrate Judge Thomas Hixson ruled on several outstanding discovery disputes, siding mostly with Meta. Judge Hixson determined that several of the plaintiffs’ discovery requests were time-barred, with the written discovery deadline having passed on October 18th. The Court also determined that the plaintiffs’ third-party subpoenas came too late and ordered the plaintiffs to notify those served after October 18th of the Court’s order. For non-written discovery, the Court also sided with Meta, explaining that the plaintiffs’ motions were filed too close to the fact discovery cut-off date, December 13th.
The plaintiffs quickly requested relief from Judge Hixson’s order. In a motion filed November 19th, the plaintiffs argued that Judge Hixson’s order was “clearly erroneous” and contrary to the case management deadline. Additionally, they explained that if the order stands, there will not be a proper record for trial. Without third-party discovery and a significant amount of production they expected Meta would be compelled to produce, the plaintiffs’ route to success seems to be narrowing. Judge Chhabria’s decision on this matter will have a significant effect on the case.
6. Sarah Andersen v. Stability AI
Background: Visual artists filed this putative class action on January 13th, 2023, alleging direct and induced copyright infringement, DMCA violations, false endorsement and trade dress claims based on the creation and functionality of Stability AI’s Stable Diffusion and DreamStudio, Midjourney Inc.’s generative AI tool, and DeviantArt’s DreamUp. On August 12th, 2024, the Court dismissed many of the claims in the plaintiffs’ first amended complaint, leaving the claims for direct copyright infringement, trademark, trade dress, and inducement. The assigned judge is Judge William H. Orrick.
Current Status: Two plaintiffs drop out. On November 21st, two plaintiffs voluntarily dismissed their DMCA claims against Stability AI. Previously, in August, the Court dismissed DMCA claims from the amended complaint, holding that DMCA 1202(b) claims require “identicality” between original works and copies with removed CMI. As reported a few weeks ago, the plaintiffs filed a second amended complaint, which included DMCA claims. The parties have now agreed that dismissing these claims from the second amended complaint, and the DMCA-plaintiffs’ withdrawal, will not prevent the plaintiffs from requesting reconsideration if the “identicality” issue is resolved in another case’s appeal.
7. Dow Jones & Co, et al v. Perplexity AI
Background: On October 21st, 2024 The Wall Street Journal and The New York Post sued generative search company Perplexity AI in the U.S. District Court for the Southern District of New York for copyright infringement and other related claims. A new twist in this litigation is the focus on Retrieval Augmented Generation (“RAG”) AI. RAG GenAI not only uses an LLM trained on copyrighted material to respond to individual prompts, but also goes out to the web to update itself based on the relevant query. Perplexity even said the quiet part out loud, encouraging its users to “skip the links” to the actual sources of the copyrighted content. Based on Perplexity’s RAG model, the media plaintiffs allege that Perplexity is infringing on their copyrights at the input and output stage, sometimes reproducing copyrighted content verbatim. Plaintiffs cited their parent company News Corp’s recent licensing agreement with OpenAI in explaining that GenAI technology can be developed by legitimate means.
Current Status: No major substantive developments this past week. Perplexity requested, and was granted, a 45-day extension to file its answer or otherwise respond to the plaintiffs’ complaint. It now has until December 30th, meaning we probably won’t have any substantive updates in this case until at least next month. In somewhat interesting news, the Court granted former Attorney General Bill Barr’s motion to practice pro hac vice in the case.
8. UMG Recordings v. Suno
Background: The RIAA on behalf of the major record labels filed their lawsuit in the federal district Court in Massachusetts on June 24th, 2024, for mass copyright infringement and related claims based on alleged training on their copyrighted works. Suno is a generative AI service that allows users to create digital music files based on text prompts. This is the first case brought against an AI service related to sound recordings. In their answer on August 1st, Suno argued that their actions were protected by fair use. The judge assigned is Chief Judge F. Dennis Saylor, IV.
Current Status: No major substantive developments this past week. The Court held a hearing on November 18th regarding the case schedule, ESI protocol, and various provisions of the protective order. On November 20th, the Magistrate Judge issued an order regarding what was discussed at the hearing. The Judge ordered the parties to disclose to each other which attorneys are permitted to receive “Highly Confidential Information.” Regarding ESI, the Court held the parties need not include on the privilege log privileged communications that post-date the complaint. This case is still very much in the early stages of discovery.
9. UMG Recordings v. Uncharted Labs (d/b/a Udio)
Background: This action was brought on June 24, 2024, in the Southern District of New York, by a group of major record companies against the company behind Udio, a generative AI service launched in April 2024 by a team of former researchers from Google Deepmind. Much like Suno above, Udio allows users to create digital music files based on text prompts or audio files. And as with the complaint against Suno (see above), plaintiffs rely on tests comprising targeted prompts including the characteristics of popular sound recordings — such as the decade of release, the topic, genre, and descriptions of the artist. They allege that using these prompts caused Udio's product to generate music files that strongly resembled copyrighted recordings. The claims are for direct infringement and related causes of action. The judge assigned is Judge Alvin K. Hellerstein.
Current Status: No major substantive developments this past week. The parties updated the Court that they have agreed on the appointment of Hon. James C. Francis IV as special master. Judge Francis previously spent thirty years as a Magistrate Judge in the Southern District of New York. Special masters can be incredibly helpful when managing discovery-intensive cases. If the Court approves this appointment, Judge Francis should help move this case along
10. Concord Music Group, et al. v. Anthropic
Background: UMG, Concord Music and several other major music companies sued Amazon-backed OpenAI competitor Anthropic on October 18th, 2023 in the U.S. District Court (Middle District of Tennessee). The music companies assert that Anthropic is infringing their music lyric copyrights on a massive scale by scraping the entire web to train its AI, essentially sucking up their copyrighted lyrics into its vortex – all without any licensing, consent or payment. In its response, Anthropic claimed fair use. The case was transferred to the Northern District of California on June 26th, 2024 and closed in Tennessee. The judge assigned is Judge Eumi K. Lee. The parties have not yet had a case management conference.
Current Status: No major substantive developments this past week. The Court held a hearing on November 21st to discuss the parties’ disputes over discovery related to Anthropic’s AI models. The plaintiffs want access to data regarding both predecessor and upcoming AI models developed by Anthropic. Anthropic claims they have already provided discovery related to several AI models, and the others requested are irrelevant. The Court should release an order regarding the issues discussed soon.
11. In re OpenAI ChatGPT Litigation (the cases, Paul Tremblay v. OpenAI, Inc., Sarah Silverman v. OpenAI, Inc., and Chabon v. OpenAI were consolidated and recaptioned to this new moniker)
Background: Comedian Sarah Silverman and other artists filed this class action lawsuit in the Northern District of California on June 28th, 2023, asserting copyright infringement claims, in addition to unfair competition, negligence, and unjust enrichment. The plaintiffs alleged that OpenAI used their copyrighted written works to train its AI chatbot. In February, the Court dismissed most of the claims against OpenAI, rejecting plaintiffs’ argument that the content generated by ChatGPT (i.e., the “output”) infringes their copyrighted works because there is no “substantial similarity” on the “output” side of the copyright question (and, therefore, no meaningful harm). But the Court gave the plaintiffs an opportunity to amend their complaint to plead a more direct link of harm (which they later did). In July, the Court dismissed the unfair competition claim. The claim for direct infringement is the only main one that remains. The case is assigned to Judge Araceli Martinez-Olguin.
Current Status: The upcoming discovery conference should provide guidance on discovery extension. As we reported last week, the parties reached an agreement on extending several case deadlines and submitted their proposal to the Court on November 8th. If approved, fact discovery will close six months after the current cut-off date, on July 25th, 2025. Other deadlines for expert reports, Daubert motions, and class certification motions would also be moved back several months. A status conference is scheduled for November 26th, so we may get an answer from the Court ahead of Thanksgiving.
12. Getty Images v. Midjourney and Stability AI
Background: Getty Images filed this lawsuit against image generator Stability AI on February 2nd, 2023, accusing the company of infringing more than 12 million photographs, their associated captions and metadata, in building and offering Stable Diffusion and DreamStudio. Getty’s claims are similar to those in The New York Times v. Microsoft & OpenAI case above, but here they are in the context of visual images instead of written articles - i.e., unlicensed scraping by their AI with an intent to compete directly with, and profit from, Getty Images (i.e., market substitution). This case also includes trademark infringement allegations arising from the accused technology’s ability to replicate Getty Images’ watermarks in the AI outputs. Getty filed its Second Amended Complaint on July 8th, 2024, and the parties are currently engaged in jurisdictional discovery related to defendants’ motion to transfer the case to the Northern District of California. The judge assigned is Judge Jennifer L. Hall.
Current Status: Still no developments. We continue to wait on Getty’s August 26th request for oral argument on the defendants’ motion to dismiss and motion to transfer. The Court has not ruled on that request, but it should be an interesting hearing if it goes forward.
13. The Authors Guild, et al. v. OpenAI
Case Background. The Authors Guild and seventeen individual authors (including John Grisham, George R.R. Martin and Nicholas A. Basbanes) filed a putative class-action suit against OpenAI on September 19th, 2023. The plaintiffs claimed that OpenAI trained its ChatGPT LLM by copying their copyrighted works. The complaint brings claims under 17 U.S.C. §501 for direct, vicarious, and contributory copyright infringement. The case is assigned to Judge Ona T. Wang.
Current Status: No major substantive developments this week. We are still waiting for the court’s decision on discovery consolidation. The case seems to be on ice until then.
14. NEW INTERNATIONAL CASE TRACKED: GEMA v. OpenAI
GEMA, a German association representing more than 95,000 composers, lyricists and publishers, filed suit in German court accusing OpenAI of reproducing their members’ song lyrics without a license. Gema claims this is a test case to clarify the law in Germany, and that it aims to establish a license model that would compensate music creators whose works are used to train AI models. The details of German copyright law are a bit beyond the scope of this blog, but we did think it noteworthy that the litigation trend is catching on worldwide. While we don’t plan to track this case closely, we will watch for any momentous developments.
Click here to read more.