Careers

Diversity

With 18 offices, from New York to Los Angeles and Seattle to Orlando, we live and work in communities all over the U.S., collaborating across locations, disciplines and even borders. Our experienced, highly ranked lawyers serve businesses wherever they are and wherever they want to go.

Offices

About Us

Pages

Professionals

Thought leadership: blogs, podcasts, videos, publications, events and news

Insights

Practices & Industries

Pro Bono

Through the ADventures in Law Blog, BakerHostetler’s Advertising, Marketing, and Digital Media team provides updates and analysis on legal issues and trends in the advertising and marketing industry.

Please visit our <a href="https://www.adventures-in-law.com/" target="_blank" rel="noopener">ADventures in Law Blog</a> site for more and to subscribe!

ADventures in Law

The Antitrust Advocate blog provides insights and commentary surrounding complex antitrust litigation and trade regulation. It’s presented by members of BakerHostetler’s Antitrust and Competition Team.

Please visit our <a href="https://www.antitrustadvocate.com/">Antitrust Advocate</a> blog site for more and to subscribe!

Antitrust Advocate

BakerHostetler’s Blockchain Monitor is a weekly blog curated and written by members of our Web3 and Digital Assets team. Published every Monday, the blog is dedicated to providing our clients, friends and industry contacts with a concise weekly update of key developments in the fast-paced blockchain industry. By subscribing below you will receive the blog in your email inbox every Monday afternoon.

Please visit our <a href="https://www.theblockchainmonitor.com/">Blockchain Monitor</a> blog site for more and to subscribe!

Blockchain Monitor

Commentary addressing risks and opportunities through the life cycle of data, technology, advertising and innovation.

Please visit our <a href="https://www.bakerdatacounsel.com/">Data Counsel</a> blog site for more and to subscribe!

Data Counsel

The Discovery Advocate blog is presented by members of BakerHostetler’s E-Discovery Advocacy and Management Team (EDAM).  Today’s modern discovery practice requires a practical approach backed by specialized expertise; Discovery Advocate examines the developments and trends that impact today’s litigants in the trenches of litigation and beyond.

Please visit our <a href="https://www.discoveryadvocate.com/">Discovery Advocate</a> blog site for more and to subscribe!

Discovery Advocate

The Employment Law Spotlight blog is written by attorneys of our national Labor and Employment Group, with an eye toward helping employers avoid litigation.

Please visit our <a href="https://www.employmentlawspotlight.com/">Employment Law Spotlight</a> blog site for more and to subscribe!

Employment Law Spotlight

BakerHostetler is at the forefront of national law firms providing clients involved in every facet of healthcare delivery across the country with comprehensive legal counsel of remarkable responsiveness, creativity, quality and value.

Please visit our <a href="https://www.healthlawupdate.com/">Health Law Update</a> blog site for more and to subscribe!

Health Law Update

In today’s global economy, patents, trademarks, copyrights, trade secrets, domain names and other intellectual property (IP) are among the most valuable assets of market leaders and rising businesses.

Please visit our <a href="https://www.ipintelligencereport.com/">IP Intelligence Report</a> blog site for more and to subscribe!

IP Intelligence Report

Updates from Capitol Hill.

Please visit our <a href="https://www.ohioclock.com/">Ohio Clock</a> blog site for more and to subscribe!

Ohio Clock

The Bargaining Table blog is written by attorneys of BakerHostetler’s Labor Relations team, with an eye toward helping employers manage their unionized workforces.

Please visit our <a href="https://www.thebargainingtableblog.com/">Bargaining Table</a> blog site for more and to subscribe!

The Bargaining Table

Am Law 100 firm with practices in Business, Digital Assets and Data Management, Intellectual Property, Labor & Employment, Litigation and Tax.

Home | BakerHostetler

2025 BHIPP - BakerHostetler IP Perspective

2025 DSIR Report Deeper Dive

32nd Annual Legislative Seminar Series

33rd Annual Legislative Seminar Series

Join us for AD Nauseam, an AdLaw podcast series featuring Amy Mudge and Daniel Kaufman, partners from our Advertising, Marketing, and Digital Media team. Together, they have decades of advertising experience and share their insights on trending legal issues in the advertising space, with an emphasis on the FTC and NAD developments.

AD Nauseam

AD-ttorneys@law

The use of noncompetes is now shifting across the country. Many states are limiting the use of noncompetes. Several states have enacted or are considering prohibitions on noncompetes. Nationally, the Federal Trade Commission is engaging in a rulemaking that could ban noncompete agreements between employers and employees. What does all of this mean for your business? Are there alternatives to noncompetes that can protect your business?

To help bring all this into focus, we’ve created a three-part series called “Alternatives to Noncompetes.”

Alternatives to Noncompetes

BakerHostetler Blockchain University

BakerHosts

BHIPP - BakerHostetler IP Perspective - 2024

Celebrating Black Voices at BakerHostetler

Since the recent Dobbs decision, which overturned Roe v. Wade, companies have been impacted nationwide and have several new legal angles to consider as it relates to their employees and their businesses. With that in mind, we’ve created a new podcast series, Dobbs on Demand, designed to help you navigate this new and evolving landscape.

Dobbs on Demand

DSIR Report Deeper Dive - 2022

DSIR Report Deeper Dive - 2023

DSIR Report Deeper Dive - 2024

Estate and Tax Planning for Globally Mobile Clients

MDLephant: What's Big in MDL Litigation?

Promoting the Progress of Science

Rev Up!

Technology from the Top: AI, Data and the Future of Business with Today’s CIOs

The Cloakroom podcast series brings you inside Capitol Hill with insightful interviews and discussions with leading lawmakers, both Republicans and Democrats from the House and Senate. Hosted by former Congressman Peter Roskam, who leads BakerHostetler’s Federal Policy team, you’ll hear firsthand about the latest policy debates on taxes, healthcare, energy, financial services and more.
Subscribe to The Cloakroom
<a href="https://podcasts.apple.com/us/podcast/the-cloakroom-with-peter-roskam-bakerhosts/id1759425935" target="_blank" rel="noopener">Apple Podcast</a> | <a href="https://www.amazon.com/music/player/podcasts/bdb139af-518d-469a-998b-ea9b49fafd93/the-cloakroom-with-peter-roskam---bakerhosts" target="_blank" rel="noopener">Amazon Music</a> | <a href="https://open.spotify.com/show/7LfpZdBC0OA2jyVjgVXNFF?si=de8f473382704494" target="_blank" rel="noopener">Spotify</a>

The Cloakroom with Peter Roskam

The Data Economy and You

The Data Stream

The Emerging New Era for Noncompetes and Trade Secrets

The Now and Next in International Trade

The State Tax Show

Wait, what?

Case Tracker: Artificial Intelligence, Copyrights and Class Actions

core/button

core/buttons

core/heading

In two nearly identical lawsuits, The Intercept Media v. OpenAI and Raw Story Media v. OpenAI, a trio of news organizations represented by the same firm alleged DMCA violations arising out of the alleged inclusion of plaintiffs’ works of journalism in the datasets used to train ChatGPT.

core/paragraph

As of April 2025, this case has been consolidated into In re: OpenAI Copyright Infringement Litigation.&nbsp;

The Intercept’s sole remaining claim is that OpenAI violated § 1202(b)(1). The court dismissed all of Raw Story Media’s claims.

Case Background

The Intercept Media, Inc. v. OpenAI, Inc., No. 1:24-cv-01515

Raw Story Media, Inc. v. OpenAI, Inc., No. 1:24-cv-01514

Court: U.S. District Court for the Southern District of New York

The Intercept Media v. OpenAI – District Judge Jed S. Rakoff and Magistrate Judge Henry J. Ricardo&nbsp;

Raw Story Media v. OpenAI – District Judge Colleen McMahon&nbsp;

Plaintiffs – The Intercept Media, Inc.; Raw Story Media, Inc.; and AlterNet Media, Inc., represented by Loevy &amp; Loevy.

OpenAI – OpenAI, Inc.; OpenAI GP, LLC; OpenAI, LLC; OpenAI OPCO LLC; OpenAI Global LLC; OAI Corporation, LLC; and OpenAI Holdings, LLC, represented by Morrison &amp; Foerster LLP and Latham &amp; Watkins LLP.

Former defendant – Microsoft Corporation, represented by Orrick, Herrington, &amp; Sutcliffe LLP

contentpilot/accordion

Status Updates

April 18, 2025 – Raw Story Media moved for reconsideration

Raw Story Media asked the court to reconsider its order to  granting OpenAI’s motion to dismiss [117] and denying plaintiffs’ motion to amend their complaint [118-119]. Plaintiffs argued that the Judicial Panel on Multidistrict Litigation granting OpenAI’s request to transfer all copyright-adjacent cases against it, including this litigation, into a multidistrict litigation (“MDL”) shortly after the court denied plaintiffs’ motion to amend their complaint [118-119] constituted a changed circumstance warranting reconsideration.  

Raw Story Media explained that all other sets of plaintiffs in the MDL survived challenges to their standing to bring § 1202(b)(1) claims against OpenAI, making them the lone outlier. Further, they emphasized OpenAI’s argument in favor of MDL that MDL was necessary to streamline discovery and “avoid inconsistent rulings on important and emerging questions of law,” highlighting the conflicting rulings on standing in Raw Story and The Intercept. Plaintiffs explained that forcing them to appeal the unfavorable court orders to the Second Circuit cuts against OpenAI’s arguments in favor of MDL, especially if the Second Circuit ultimately reversed.

On the standing issue, plaintiffs argued that, like the other plaintiffs involved in the MDL, they suffered harm analogous to copyright infringement through OpenAI downloading tens of thousands of their copyrighted works and intentionally removing CMI before using the CMI-less copies to train OpenAI’s GPT models.&nbsp;

Finally, plaintiffs argued that they should be granted leave to amend their complaint because OpenAI had not argued that permitting them to do so would cause OpenAI any prejudice and only opposed based on the supposed inadequacy of the pleadings. Plaintiffs further contended that they had statutory standing under § 1203(a) and had plausibly alleged intentional removal of CMI by OpenAI through the Dragnet and Newspaper algorithms as well as that OpenAI had reasonable grounds to know that such CMI removal would induce, enable, facilitate, or conceal copyright infringement, as OpenAI was aware that ChatGPT outputs could infringe copyrights.

April 18, 2025 – The Intercept filed its second amended complaint

The Intercept filed a second amended complaint seeking damages and injunctive relief based on OpenAI’s alleged violations of § 1202(b)(1). The Intercept also included allegations that Microsoft violated § 1202(b)(1) and both defendants violated § 1202(b)(3) to preserve its appellate rights, which were similar to allegations supporting those claims in its first amended complaint [87].&nbsp;

The Intercept alleged that OpenAI used two algorithms—Dragnet and Newspaper—to extract text, including The Intercept articles, from websites for inclusion in its training datasets, and knew that these algorithms did not extract CMI accompanying the extracted text to create WebText and WebText2. In support, The Intercept attached the results of a data scientist’s application of the two algorithms to URLs associated with The Intercept’s copyrighted articles allegedly included in OpenAI’s training datasets, which contained substantively identical text but omitted CMI. The Intercept also alleged that a recreation of Common Crawl included thousands of URLs associated with its copyrighted articles, again omitting CMI. 

Moreover, The Intercept alleged that OpenAI’s removal of CMI both concealed OpenAI’s own copyright infringement in the process of training ChatGPT and facilitated downstream copyright infringement in ChatGPT outputs generated and further distributed by end users. The Intercept claimed that OpenAI had publicly admitted their knowledge that ChatGPT incorporates and sometimes even regurgitates verbatim copyrighted works of journalism in outputs without associated CMI, pointing to recent changes OpenAI made to ChatGPT reducing the likelihood of these occurrences. The Intercept included several examples of ChatGPT outputs containing regurgitations of its copyrighted articles without CMI.&nbsp;

April 3, 2025 – Court denied Raw Story Media’s motion for leave to amend.

The court denied Raw Story Media’s motion for leave to amend its complaint [118-119], finding that the allegations added to plaintiffs’ proposed amendments did not confer standing to pursue either damages or injunctive relief. With respect to damages, the court found that plaintiffs still failed to plausibly allege that a copy of any of their copyrighted works with CMI removed was disseminated by ChatGPT to anyone in response to a specific inquiry. The court explained that plaintiffs’ new allegations on the technical process by which OpenAI uploaded their copyrighted works into OpenAI’s training datasets did not render plaintiffs’ asserted harm more concrete.  

Moreover, the court found that plaintiffs still lacked standing to pursue injunctive relief because the likelihood of ChatGPT outputting plagiarized content from plaintiffs’ copyrighted works seemed remote. The court explained that plaintiffs’ reliance on allegations of plagiarism made by other litigants in other lawsuits against OpenAI did not show “certainly impending” injury or “substantial risk” that they will suffer the same harm.&nbsp;

February 20, 2025 – Court granted Microsoft’s and (in part) OpenAI’s motion to dismiss The Intercept’s complaint.

OpenAI’s Motion to Dismiss. The court granted OpenAI’s motion to dismiss The Intercept’s § 1202(b)(3) claim but denied OpenAI’s request to dismiss the § 1202(b)(1) claim.&nbsp;&nbsp;

Regarding the § 1202(b)(3) claim, the court found that The Intercept’s complaint lacked factual support for its allegation that OpenAI distributed its articles. Regarding OpenAI’s alleged sharing of the ChatGPT training dataset (including The Intercept’s articles) with Microsoft, the court explained that “the mere fact that Microsoft has a partnership with OpenAI and that it provides a data center and supercomputing infrastructure” was an insufficient indication that the companies shared training data for their competing AI products or that they shared The Intercept’s articles in particular. 

As for the § 1202(b)(1) claim, the court found that The Intercept had plausibly alleged OpenAI’s intentional removal of CMI from its articles through its use of the Dragnet and Newspaper algorithms. Further, the court found that The Intercept had plausibly alleged that OpenAI had reasonable grounds to know that its removal of CMI would induce, enable, facilitate, or conceal copyright infringement. Although the court rejected the theory that OpenAI had the requisite knowledge that its removal of CMI would conceal its own infringement in reproducing The Intercept’s articles in training datasets, it found persuasive a second theory that OpenAI’s removal of CMI would facilitate downstream infringement through regurgitations of its articles in ChatGPT outputs. It reasoned that The Intercept had plausibly alleged that OpenAI knew ChatGPT regurgitated verbatim copyrighted-protected works of journalism in outputs (including some of The Intercept’s works), as well as that some users would likely infringe copyrights through ChatGPT outputs.&nbsp;&nbsp;

Microsoft’s Motion to Dismiss. The court granted Microsoft’s motion to dismiss The Intercept’s § 1202(b) claims.

Regarding § 1202(b)(1), the court found that none of the evidence related to Microsoft’s involvement with ChatGPT, including its partnership with and 49% ownership stake in OpenAI, its alleged provision of the data center and supercomputing infrastructure used to train ChatGPT, and its CEO’s statement that “we have the data,” bore any relationship to the alleged CMI removal by OpenAI.   

Moreover, the court dismissed The Intercept’s § 1202(b)(3) claim against Microsoft for the same reasons it dismissed the analogous § 1202(b)(3) claim against OpenAI.&nbsp;

Dec. 20, 2024 – OpenAI opposed motion to amend Raw Story Media complaint

OpenAI argued that plaintiffs’ proposed amendments are futile. OpenAI argued that plaintiffs are not “persons injured” within the meaning of DMCA.

Dec. 5, 2024 – OpenAI answered The Intercept complaint

OpenAI filed an answer to The Intercept’s amended complaint [87]. OpenAI denied the single remaining cause of action – violation of 17 U.S.C. §1202(b)(1).

Nov. 21, 2024 – Plaintiffs requested leave to file amended complaint in Raw Story Media

Raw Story Media and AlterNet Media filed a motion for leave to amend its complaint or alternatively to continue jurisdictional discovery regarding the defendants’ dissemination of plaintiffs’ works. Plaintiffs argued that the proposed amendments present a close analogy between copyright infringement and Plaintiffs’ DMCA injuries based on the specific facts of this case. Plaintiffs also argued their proposed amendments justify standing based on disgorgement of unlawful profits. Plaintiffs argued that its proposed amended complaint would survive a 12(b)(6) challenge.

Nov. 21, 2024 – Court dismissed all but one claim in The Intercept

The court dismissed with prejudice all of The Intercept’s claims against Microsoft and its §1202(b)(3) claim against OpenAI. The court permitted The Intercept’s&nbsp;&nbsp; §1202(b)(1) claim against OpenAI to proceed.

Nov. 7, 2024 – Court granted OpenAI’s motion to dismiss Raw Story Media complaint.

The Court agreed with defendants that the Raw Story Media and AlterNet Media Inc. lacked Article III standing to assert their claims. The plaintiffs were found to lack a concrete injury-in-fact.

Aug. 27, 2024 – The Intercept filed notice of supplemental authority ISO its opposition to OpenAI’s motion to dismiss

The Intercept filed a notice of supplemental authority regarding the Second Circuit’s decision in Michael Grecco Prods., Inc. v. RADesign, Inc. No. 23-1078, 2024 WL 3836578 (2d Cir. Aug. 16, 2024), which The Intercept argued supported its position in its opposition [90] to OpenAI’s motion to dismiss [89] that its DMCA § 1202(b) claims were not barred by the statute of limitations.&nbsp;&nbsp;

The Intercept explained that the Second Circuit had clarified that, under the discovery rule, the sophistication of a plaintiff did not automatically relieve the defendant of its burden to plead and prove a Copyright Act statute of limitations defense. In addition, it explained that the Second Circuit reiterated that, for dismissal on limitations grounds, it must be “clear from the face of the complaint” that the plaintiff, exercising due diligence, would have discovered its claims less than 3 years before it filed the lawsuit. Finally, it explained that the Second Circuit had emphasized that a defendant’s statute of limitations argument is an affirmative defense, so plaintiffs need not allege timeliness of their claims in the complaint.

Aug. 22, 2024 – OpenAI filed notice of supplemental authority ISO its motion to dismiss in each case

OpenAI filed a notice of supplemental authority regarding a court order in <a href="https://www.bakerlaw.com/andersen-v-stability-ai/" target="_blank" rel="noreferrer noopener">Andersen v. Stability AI</a>, which granted defendants’ motions to dismiss plaintiffs’ DMCA § 1202(b) claims. OpenAI argued the court order supported its motions to dismiss [68 &amp; 89] the complaints filed by Raw Story Media [1] and The Intercept Media [87]. OpenAI explained that the Andersen court held plaintiffs’ § 1202(b) claims failed “[b]ecause there are no allegations that any output from [the models] was identical to a plaintiff’s work,” an argument OpenAI had asserted in its motions to dismiss each complaint.&nbsp;

July 15, 2024 – The Intercept filed a supplemental brief opposing Microsoft’s and OpenAI’s motions to dismiss

The Intercept filed a supplemental brief in opposition to Microsoft’s [88] and OpenAI’s [89] renewed motions to dismiss its amended complaint [87]. Regarding standing, The Intercept argued that it had sufficiently alleged a concrete injury based on the defendants’ “interference with its copyright-protected property,” which was analogous to copyright infringement claims for which public dissemination is not required. However, if the court agreed with defendants that public dissemination of CMI-less copies of its works was required, it argued it was entitled to jurisdictional discovery.

As for OpenAI’s statute of limitations argument, The Intercept contended that it “has plausibly alleged that OpenAI employed the same CMI removal processes,” web scraping algorithms designed to remove CMI from news articles, “within the last three years.” It also argued that the discovery rule, if applied, may extend the applicable statute of limitations period.

Regarding the DMCA’s double-scienter requirement, The Intercept argued that defendants had reason to know their removal of CMI would (1) conceal their own infringement during training from users by omitting CMI in outputs; (2) conceal the infringing nature of outputs from users; and (3) induce, enable, and facilitate users to infringe by further distributing infringing outputs to others, as some users would be less likely to do so if they knew the outputs were infringing. And in response to defendants’ argument that ordinary users would never produce the regurgitations attached to its complaint, it emphasized that “ChatGPT is far less prone to regurgitation now than it was years ago.”

The Intercept also argued it had plausibly alleged its § 1202(b)(3) claims by including allegations that defendants “distributed the training data, from which CMI had been removed, with each other.” It pointed to an interview quote from Microsoft’s CEO that “we have the data,” leading to a “plausible inference” that Microsoft and OpenAI exchanged the training data. Further, it argued that defendants had reason to know such distribution of CMI-less copies would conceal their infringement, as “clearly the recipient of an unattributed wall of text … would have no idea that it was an unauthorized copy of The Intercept’s works.”

Finally, The Intercept rebutted Microsoft’s arguments regarding its involvement in creating the training datasets at issue by relying upon allegations concerning Microsoft’s admission that “we have the data” as well as its “close partnership” with OpenAI and provision of bespoke supercomputing resources used to power ChatGPT.

July 8, 2024 – Microsoft and OpenAI filed supplemental briefs ISO their motions to dismiss The Intercept’s amended complaint

Microsoft. Microsoft filed a supplemental brief in support of its motion to dismiss The Intercept’s amended complaint [87]. Microsoft first argued that The Intercept does not have standing because ChatGPT and Copilot “have never generated, and no facts plausibly suggest they are likely ever to generate, outputs matching The Intercept’s copyrighted works” under “realistic conditions.” As a result, “there is no public dissemination of a CMI-less copy of The Intercept’s works; [and] without public dissemination, there is no concrete injury.” And even assuming Microsoft shared CMI-less copies with OpenAI, “dissemination between the Defendants is not the sort of public lack of attribution that may cause injury.”

Second, Microsoft contended that it was not involved in creating the datasets used to train ChatGPT and Copilot. It explained that The Intercept’s allegations related to OpenAI’s use of “Dragnet and Newspaper” algorithms to scrape copyrighted articles when creating the WebText datasets were aimed at OpenAI, not Microsoft. Further, The Intercept’s reliance on “Microsoft’s general collaboration with OpenAI and provision of computing and cloud technology” was “a legally insufficient basis” for a § 1202(b) claim.

Finally, Microsoft argued that The Intercept did not plausibly establish § 1202(b)’s double-scienter requirement, explaining “[i]t is objectively implausible that removal of CMI could ever facilitate or conceal copyright infringement because the GPT-based tools are unlikely to ever reproduce The Intercept’s content at all.” Further, Microsoft’s lack of involvement in creating the datasets at issue prevented it from having “any reason to believe that removal of CMI would facilitate piracy” of The Intercept’s works.

OpenAI. OpenAI likewise filed a supplemental brief, characterizing The Intercept’s lawsuit as an improper attempt to use DMCA § 1202(b) as “a back-door to a claim for copyright infringement” because The Intercept failed to register its copyrighted works prior to the alleged infringement and thus could not pursue statutory damages. It emphasized that “every court to have considered whether a Section 1202 claim can be pled” based on similar conduct had concluded that it cannot, citing <a href="https://www.bakerlaw.com/the-copilot-litigation/">Doe v. GitHub, Inc</a>. and <a href="https://www.bakerlaw.com/kadrey-v-meta/">Kadrey v. Meta</a>.

Like Microsoft, OpenAI argued that The Intercept lacks standing because “the exclusion of CMI from internal datasets that never see the light of day is not, standing alone, a concrete injury that could satisfy Article III’s requirement.” As for the alleged “regurgitations” of CMI-less copies of The Intercept’s works generated by ChatGPT, OpenAI argued that “The Intercept does not allege that any ChatGPT user—other than The Intercept—has ever used similar prompts or otherwise caused ChatGPT to output portions” of such works.

Regarding the merits, OpenAI first argued that The Intercept’s § 1202(b)(1) claim was time-barred because it “hinge[d] on an alleged act of CMI removal” in connection with the WebText dataset “that occurred five years before it chose to file this action,” outside of the applicable three-year statute of limitations.

Moreover, OpenAI pointed out a lack of allegations that plausibly establish § 1202(b)’s double-scienter requirement. OpenAI explained that its alleged removal of CMI in internal datasets could not conceal infringement because “the absence of CMI in a non-public dataset cannot have any effect on what the public does or does not know.” As for CMI-less ChatGPT outputs, OpenAI explained that The Intercept failed to allege that OpenAI had “reasonable grounds to know” that (1) ChatGPT would regurgitate any of The Intercept’s articles at the time of such CMI removal or (2) that “its alleged exclusion of CMI from The Intercept’s articles in training datasets had any ‘identifiable connection’ to the inclusion or exclusion of CMI in potential ChatGPT outputs.”

Finally, OpenAI challenged The Intercept’s § 1202(b)(3) claim as “conclusory,” noting the absence of specific allegations as to “when, why, or how” OpenAI “shared” identical but CMI-less copies of The Intercept’s works with Microsoft.

June 21, 2024 – The Intercept filed amended complaint

The Intercept amended its complaint against OpenAI and Microsoft in an attempt to address the lack of specificity in its original complaint. The amended complaint retained both DMCA claims asserted against each defendant:

core/list

The amended complaint attempted to add specificity by explaining the process by which defendants curated and cleaned their LLM training datasets. It explained that OpenAI admitted to using “Newspaper” and “Dragnet” web scraping algorithms to develop the WebText dataset used to train ChatGPT and how those algorithms removed CMI from copies that were “substantively identical” to the original articles. The Intercept also detailed how scientists created an “approximation” of the WebText dataset, which it relied on to identify which of its articles had been included.

Moreover, the amended complaint added details alleging the tendency of ChatGPT and Copilot to regurgitate copyrighted works without CMI in its outputs. It asserted that OpenAI possessed “a repository of every regurgitation of [p]laintiff’s works” due to OpenAI’s retention of users’ chat histories with ChatGPT. It also claimed that OpenAI has “effectively admitted that regurgitation of copyrighted works is infringement,” relying on copyright-related error messages received from ChatGPT when attempting to obtain the same regurgitations attached to the complaint in <a href="https://www.bakerlaw.com/daily-news-v-microsoft/">Daily News v. Microsoft</a>.

As for the DMCA’s double scienter requirement, The Intercept relied on OpenAI’s creation of opt-opt tools for copyright owners, as well as each defendant’s pledge to indemnify ChatGPT and Copilot users against copyright infringement claims, to allege that defendants had reason to know their removal of CMI and distribution of CMI-less copies would induce, enable, facilitate, or conceal copyright infringement.

The amended complaint also attempted to explain the relationship between Microsoft and OpenAI in greater detail.

Finally, the amended complaint included more than 600 pages of exhibits, including a list of The Intercept’s copyrighted works as well as a list describing which of those works were included in an approximation of the WebText dataset.

June 6, 2024 – Order granting leave for The Intercept to amend its complaint

After considering the parties’ briefings and oral argument on defendants’ motions to dismiss, the Court issued an order granting leave for The Intercept to file an amended complaint “to attempt to rectify some of the seeming lack of specificity in its current complaint.” It ordered The Intercept to file an amended complaint by June 21, 2024.

The Court also explained that defendants may file a supplemental brief in support of their motions to dismiss by July 8, 2024, and The Intercept may then file a supplemental brief in opposition by July 15, 2024.

May 20, 2024 – OpenAI replied to Raw Story Media and AlterNet Media’s opposition to its motion to dismiss

In OpenAI’s reply [71] in support of its motion to dismiss [68], it maintained that the plaintiffs lacked standing because they had failed to identify any of their copyrighted works from which OpenAI had allegedly removed CMI, continuing to rely on <a href="https://www.bakerlaw.com/the-copilot-litigation/">Doe 1 v. GitHub, Inc</a>. for support. OpenAI also argued that plaintiffs “are not similarly situated” to the <a href="https://www.bakerlaw.com/the-copilot-litigation/">Doe 1 v. GitHub, Inc</a>. plaintiffs regarding standing for injunctive relief because “they have not shown any substantial risk that their copyrighted works will be outputted by ChatGPT in the future.”

OpenAI also maintained that plaintiffs’ complaint [1] should be dismissed for failure to state a plausible DMCA claim even if they are found to have standing. It argued that plaintiffs’ reliance on CMI-less copies of their works in “approximations of ChatGPT’s training data” to plead OpenAI’s intentional removal of CMI “at most … is merely consistent with, but does not plausibly suggest, a Section 1202 violation.” OpenAI also challenged plaintiffs’ reliance on a Copyleaks study to claim extensive plagiarism of online content in ChatGPT outputs, arguing plaintiffs’ claim was “a far cry from an allegation that OpenAI had reason to know that allegedly removing CMI would induce, enable, facilitate, or conceal ... regurgitation of [p]laintiffs’ content” in ChatGPT outputs.

Finally, OpenAI noted that “[c]ourts have repeatedly rejected claims brought under Section 1202 based on the training of AI models,” citing <a href="https://www.bakerlaw.com/openai-chatgpt-litigation/" target="_blank" rel="noreferrer noopener">Tremblay v. OpenAI</a>, <a href="https://www.bakerlaw.com/the-copilot-litigation/" target="_blank" rel="noreferrer noopener">Doe 1 v. GitHub, Inc</a>., and <a href="https://www.bakerlaw.com/andersen-v-stability-ai/" target="_blank" rel="noreferrer noopener">Andersen v. Stability AI</a>.

May 16, 2024 – Microsoft and OpenAI replied to The Intercept’s opposition of defendants’ motions to dismiss

Microsoft’s Reply. In Microsoft’s reply [79], in support of its motion to dismiss, Microsoft argued that The Intercept’s opposition [78] confirmed its DMCA claims “are as threadbare as they appear.” Microsoft insisted that The Intercept lacked Article III standing because its asserted injury is nothing more than alleged removal of CMI “during the entirely internal training and development of an LLM,” and objected to The Intercept’s proposed analogy between a DMCA claim and a copyright infringement claim.

Microsoft also maintained that The Intercept failed to state a plausible DMCA claim, calling its claims “bare conjecture based on virtually no plausible factual allegations that Microsoft has done anything” to its copyrighted works.

OpenAI’s Reply. In its reply [80], OpenAI, like Microsoft, maintained that The Intercept lacked standing to assert its DMCA claims, as it did not plausibly allege that CMI-less copies of its own works appeared in ChatGPT outputs and “cannot avoid the conclusion that ‘the mere existence of inaccurate information in a database [without dissemination] is insufficient to confer Article III standing,’” citing TransUnion LLC v. Ramirez. OpenAI also challenged The Intercept’s attempt to distinguish <a href="https://www.bakerlaw.com/the-copilot-litigation/">Doe 1 v. GitHub, Inc</a>., emphasizing that “the court rejected standing because [those] plaintiffs did not allege that they, personally, had suffered any injury.”

As for the § 1202(b)(3) claim, OpenAI argued that The Intercept failed to allege that OpenAI had distributed identical copies of its works with CMI removed and failed to adequately plead the DMCA’s double-scienter requirement, as it “made no connection between sharing data with Microsoft and concealing infringement from the public.”&nbsp;

May 13, 2024 – Raw Story Media and AlterNet Media filed an opposition to OpenAI’s motion to dismiss

In their opposition [70] to OpenAI’s motion to dismiss [68], Raw Story Media and AlterNet Media challenged OpenAI’s contention that they lacked standing to bring their DMCA claim. Plaintiffs argued that, like copyright infringement, the concrete and particularized injury they suffered was OpenAI’s “interference with [their] right to exclude others from using [their] copyrighted works” by removing CMI from them, “irrespective” of dissemination or economic harm. Plaintiffs distinguished <a href="https://www.bakerlaw.com/the-copilot-litigation/">Doe 1 v. GitHub, Inc</a>. by explaining that, unlike here, those plaintiffs characterized their injury “as a violation of their licenses, which prohibited dissemination without CMI.” They also emphasized that those plaintiffs were found to have standing to seek an injunction, despite a failure to allege dissemination of their own works, due to a “substantial risk” that their works would be reproduced in AI outputs without CMI.

Moreover, plaintiffs contended that they had plausibly stated a § 1202(b)(1) claim. They argued that they had properly identified the copyrighted works at issue and the types of CMI removed from them. Plaintiffs also noted that OpenAI’s efforts to conceal the works included in its ChatGPT training datasets prevented them from providing names for the works and argued that requiring them to do so at the motion to dismiss stage “would perversely incentivize defendants to conceal their DMCA violations.”

Plaintiffs also argued they had adequately pled the DMCA’s double-scienter requirement. They challenged OpenAI’s argument that their claims relied on “speculat[ive]” inferences about the relationship between ChatGPT’s training data and its outputs, concluding that “CMI-less regurgitations of some news articles [in outputs] means removal of CMI from all news articles [in the training data], including those owned by [p]laintiffs.” Further, plaintiffs argued that OpenAI knew or had reason to know that its removal of CMI from plaintiffs’ works would conceal OpenAI’s infringements of those works by failing to communicate to ChatGPT users that outputs were infringing.

Further, they argued that OpenAI had sufficient knowledge that such removal and concealment of its infringement would make ChatGPT users more likely to distribute infringing outputs, noting that OpenAI “promote[s] ChatGPT as a tool that can be used to generate content for a future audience.” Finally, plaintiffs emphasized that future infringement via ChatGPT outputs was likely, relying on a study that found extensive plagiarism of online content in ChatGPT outputs.

May 6, 2024 – The Intercept Media opposed Microsoft’s and OpenAI’s motions to dismiss

In its opposition [79] to defendants’ motions to dismiss, The Intercept challenged the defendants’ argument that it lacked standing. The Intercept asserted that it suffered “concrete” injuries because defendants’ unlawful removal of CMI from its copyrighted works and distribution of CMI-less copies of those works were “analogous to copyright infringement” and that dissemination of its works without CMI is not a requirement for standing to assert a DMCA claim. The Intercept distinguished the <a href="https://www.bakerlaw.com/the-copilot-litigation/">Doe 1 v. GitHub, Inc</a>. court’s dismissal of those plaintiffs’ claims for lack of standing to seek damages and explained that the court had found standing to sue for an injunction based on allegations of “a substantial risk that [d]efendants’ programs will reproduce [p]laintiffs’ licensed code as output.”

The Intercept argued that it had plausibly stated a claim under § 1202(b)(1), as it adequately alleged the existence and defendants’ removal of CMI from its works, including author, title, copyright notice, and terms of use information. The Intercept noted that the “close working relationship” between the defendants makes it plausible that Microsoft had actively participated in the removal of CMI from the works included the ChatGPT training datasets.

The Intercept also asserted it had sufficiently identified which of its copyrighted works were at issue by identifying them as those “contained in [d]efendants’ training sets,” noting that the defendants know the contents of their own training sets and the defendants’ secrecy regarding the contents of those datasets prevents it from providing names for the works.

Regarding the DMCA’s double-scienter requirement, The Intercept argued that the defendants knew or had reason to know that their removal of CMI would “induce, enable, facilitate, or conceal an infringement” because such removal would conceal the defendants’ own infringement to ChatGPT users, making them more likely to distribute infringing outputs. It also noted that OpenAI’s licensing deals with media entities and creation of opt-out tools for copyright owners evidenced OpenAI’s knowledge that use of copyrighted works to train LLMs and regurgitation of those works in outputs constitutes infringement.

Finally, The Intercept argued that it had plausibly stated a § 1202(b)(3) claim, arguing the “close working relationship” between the defendants in the development of ChatGPT makes it plausible that they shared CMI-less copies of The Intercept’s works with each other. Further, it argued that the defendants “distributing the works amongst themselves, with CMI removed, both conceals the sender’s infringement and facilitates the recipient’s [infringement] by suggesting that the works are not copyrighted and thus may be lawfully copied and distributed further.”

Apr. 29, 2024 – OpenAI moved to dismiss Raw Story Media and AlterNet Media’s complaint

OpenAI moved to dismiss Raw Story Media and AlterNet Media’s complaint [1] under Rule 12(b)(1) for lack of subject matter jurisdiction, arguing that plaintiffs failed to plead that they suffered a cognizable harm and thus lack standing. OpenAI explained that “the allegation that [p]laintiffs’ works merely appear in the training data without CMI, in the absence of an allegation that ChatGPT has actually outputted their works,” is not sufficient to demonstrate the concrete and particularized injury-in-fact needed to confer standing. OpenAI analogized to <a href="https://www.bakerlaw.com/the-copilot-litigation/">Doe 1 v. GitHub, Inc.</a>, where the court found that those plaintiffs lacked standing to assert DMCA claims against OpenAI for similar reasons.

OpenAI also moved to dismiss under Rule 12(b)(6), arguing that plaintiffs’ allegations failed to state a claim under Section 1202(b)(1) of the DMCA for two additional reasons. First, OpenAI argued that plaintiffs failed to identify any of their copyrighted works from which CMI was removed or altered, instead making only “general allegations about [their] ‘works of journalism’ that were ‘published on the internet.’”

Second, OpenAI argued that plaintiffs did not adequately plead the DMCA’s “double-scienter” requirement. OpenAI explained that plaintiffs did not sufficiently allege that OpenAI intentionally removed or altered CMI from their copyrighted works, noting that other courts had dismissed similar allegations that “the training process does not preserve CMI by design” as “conclusory.” Additionally, OpenAI contended that plaintiffs did not sufficiently allege how OpenAI had “reason to know” that any omission of CMI from copyrighted works in its training dataset could “induce, enable, facilitate, or conceal” copyright infringement. It explained that “because the CMI was allegedly removed from an internal dataset, that alleged omission has no effect on the public at all.” Further, OpenAI argued that plaintiffs’ concession that “there are ‘instances in which author or title information is included in a response’ from ChatGPT … forecloses any argument that the alleged ‘removal’ of CMI … was intended to facilitate or conceal the creation of supposedly infringing ChatGPT outputs.”

Apr. 15, 2024 – Microsoft and OpenAI filed separate motions to dismiss

Microsoft’s Motion to Dismiss. Microsoft moved to dismiss The Intercept’s complaint [1] under Rule 12(b)(1) and Rule 12(b)(6). Microsoft alleged The Intercept lacked Article III standing because it did not suffer a “concrete injury” from the “alleged removal of CMI in the non-public setting of training.” It explained that “[t]he mere presence of a copy of a work without a plaintiff’s CMI, absent dissemination, is precisely the sort of vaporous technical harm that the Supreme Court has found insufficient to confer standing.”

Microsoft argued that The Intercept failed to adequately plead either a § 1202(b)(1) or § 1202(b)(3) violation. Microsoft argued The Intercept failed to offer examples of which copyrighted works were included in its training datasets or any instance where ChatGPT or Bing AI products generated text identical to The Intercept’s copyrighted works. Instead, Microsoft characterized the complaint as relying on a string of conclusory allegations—that “thousands” of unidentified copyrighted works owned by The Intercept were included in Microsoft’s training datasets, that ChatGPT outputs “generally” do not include CMI, and that ChatGPT would have learned to communicate CMI if it was trained on works including CMI—to improperly conclude that CMI must have been removed from every work contained in Microsoft’s training datasets. Moreover, Microsoft argued that The Intercept failed to sufficiently plead the DMCA’s “double scienter” requirement, as the removal of CMI from copyrighted works in the non-public setting of model training, absent dissemination of copies without CMI, was “highly unlikely” to induce, facilitate, or conceal infringement.

For the § 1202(b)(3) claim, Microsoft argued that The Intercept’s allegations that Microsoft “created” Common Crawl datasets and provided the datacenters and computing infrastructure used to develop ChatGPT did not permit an inference that it “distributed” training datasets knowing they contained copyrighted works without CMI.

OpenAI’s Motion to Dismiss. OpenAI also moved to dismiss The Intercept’s complaint [1] under Rule 12(b)(1) and Rule 12(b)(6). Like Microsoft, OpenAI argued that The Intercept lacked standing because the complaint failed to plead a cognizable injury suffered by The Intercept, echoing the arguments made by Microsoft in its motion to dismiss. OpenAI also analogized to <a href="https://www.bakerlaw.com/the-copilot-litigation/">Doe 1 v. GitHub, Inc</a>., where the court dismissed DMCA claims for lack of standing where the plaintiffs failed to allege that their own copyrighted computer code appeared in AI model outputs.&nbsp;&nbsp;

Like Microsoft, OpenAI argued that The Intercept’s DMCA claims were insufficient under Rule 12(b)(6) because The Intercept failed to identify copyrighted works from which CMI had allegedly been removed during model training.

OpenAI argued that The Intercept did not adequately plead the DMCA’s “double scienter” requirement for § 1202(b)(1) claim. According to OpenAI, the complaint relied on “unwarranted deductions of fact” to allege that OpenAI intentionally removed CMI. OpenAI also explained that The Intercept “admits that OpenAI’s training dataset is not publicly accessible,” and therefore “omission in the training dataset alone … could not deprive the public of [CMI]” and induce, facilitate, or conceal copyright infringement.

For the § 1202(b)(3) claim, OpenAI argued that The Intercept failed to sufficiently allege “when, why, or how” OpenAI “shared” identical copies of its works without CMI with Microsoft or that OpenAI did so with reason to know that such “shar[ing]” would induce, facilitate, or conceal copyright infringement.

Feb. 28, 2024 – The Intercept Media filed complaint against OpenAI and Microsoft

In its complaint against OpenAI and Microsoft, The Intercept Media alleged that OpenAI and Microsoft violated the DMCA by (1) intentionally removing CMI from copies of its works of journalism when creating or adapting the WebText, WebText2, and Common Crawl datasets that they used to train their LLMs, and (2) sharing copies of its works of journalism from which CMI had been removed with one another as part of their efforts to develop ChatGPT. The complaint alleged two causes of action against each defendant:

In addition to damages, The Intercept seeks an injunction requiring OpenAI and Microsoft to remove all copies of its copyrighted works that do not include CMI from their training datasets and any other repositories.

Feb. 28, 2024 – Raw Story Media and AlterNet Media filed complaint against OpenAI

In their complaint against OpenAI, Raw Story Media and AlterNet Media alleged that OpenAI violated the DMCA by choosing to train ChatGPT with training datasets that included copies of their copyrighted works of journalism from which CMI had been removed. In doing so, the plaintiffs alleged that OpenAI “trained ChatGPT not to acknowledge or respect copyright, not to notify ChatGPT users when the responses they received were protected by journalists’ copyrights, and not to provide attribution when using the works of human journalists.”

The complaint alleges a single cause of action against OpenAI for violation of § 1202(b)(1), alleging that OpenAI intentionally removed CMI from copies of plaintiffs’ works included in OpenAI’s training datasets. The plaintiffs pointed to OpenAI’s creation of tools “in late 2023 to allow copyright owners to block their work from being incorporated into training datasets,” to establish scienter.

In addition to damages, plaintiffs seek an injunction requiring OpenAI to remove all copies of their works from which CMI had been stripped from OpenAI’s training datasets and any other repositories.

Select Court Documents

Feb. 28, 2024 – <a href="https://admin.bakerlaw.com/wp-content/uploads/2024/03/ECF-1-The-Intercept-Media-Complaint-No.-124-cv-01515.pdf" target="_blank" rel="noreferrer noopener">ECF 1 The Intercept Media Complaint (No. 1:24-cv-01515)</a>

Feb. 28, 2024 – <a href="https://admin.bakerlaw.com/wp-content/uploads/2024/03/ECF-1-Raw-Story-Media-Complaint-No.-124-cv-01514.pdf" target="_blank" rel="noreferrer noopener">ECF 1 Raw Story Media Complaint (No. 1:24-cv-01514)</a>

Apr. 15, 2024 – <a href="https://admin.bakerlaw.com/wp-content/uploads/2024/04/ECF-49-50-Microsofts-Motion-to-Dismiss.pdf" target="_blank" rel="noreferrer noopener">ECF 49 &amp; 50 Microsoft’s Motion to Dismiss The Intercept Complaint </a>

Apr. 15, 2024 – <a href="https://admin.bakerlaw.com/wp-content/uploads/2024/04/ECF-52-53-OpenAIs-Motion-to-Dismiss.pdf" target="_blank" rel="noreferrer noopener">ECF 52 &amp; 53 OpenAI’s Motion to Dismiss The Intercept Complaint</a>

Apr. 29, 2024 – <a href="https://admin.bakerlaw.com/wp-content/uploads/2024/05/ECF-68-OpenAIs-Motion-to-Dismiss-Raw-Story-Medias-Complaint.pdf" target="_blank" rel="noreferrer noopener">ECF 68 OpenAI’s Motion to Dismiss Raw Story Media and AlterNet Media’s Complaint</a>

May 6, 2024 – <a href="https://admin.bakerlaw.com/wp-content/uploads/2024/05/ECF-78-The-Intercepts-Opposition-to-Microsofts-and-OpenAIs-Motions-to-Dismiss.pdf" target="_blank" rel="noreferrer noopener">ECF 78 The Intercept’s Opposition to Microsoft’s and OpenAI’s Motions to Dismiss</a>

May 13, 2024 – <a href="https://admin.bakerlaw.com/wp-content/uploads/2024/05/ECF-70-Raw-Story-Media-and-AlterNet-Medias-Opposition-to-OpenAIs-Motion-to-Dismiss.pdf" target="_blank" rel="noreferrer noopener">ECF 70 Raw Story Media and AlterNet Media’s Opposition to OpenAI’s Motion to Dismiss</a>

May 16, 2024 – <a href="https://admin.bakerlaw.com/wp-content/uploads/2024/05/ECF-79-Microsofts-Reply-ISO-Motion-to-Dismiss-The-Intercepts-Complaint.pdf" target="_blank" rel="noreferrer noopener">ECF 79 Microsoft’s Reply ISO Motion to Dismiss The Intercept’s Complaint </a>

May 16, 2024 – <a href="https://admin.bakerlaw.com/wp-content/uploads/2024/05/ECF-80-OpenAIs-Reply-ISO-Motion-to-Dismiss-The-Intercepts-Complaint.pdf" target="_blank" rel="noreferrer noopener">ECF 80 OpenAI’s Reply ISO Motion to Dismiss The Intercept’s Complaint</a>

May 20, 2024 – <a href="https://admin.bakerlaw.com/wp-content/uploads/2024/05/ECF-71-OpenAIs-Reply-ISO-Motion-to-Dismiss-Raw-Story-Media-and-AlterNet-Medias-Motion-to-Dismiss.pdf" target="_blank" rel="noreferrer noopener">ECF 71 OpenAI’s Reply ISO Motion to Dismiss Raw Story Media and AlterNet Media’s Complaint</a>

<a href="https://admin.bakerlaw.com/wp-content/uploads/2024/06/ECF-81-Order-Granting-Leave-to-Amend-Complaint.pdf" target="_blank" rel="noreferrer noopener">June 6, 2024 – ECF 81 Order Granting Leave to Amend Complaint</a>

<a href="https://admin.bakerlaw.com/wp-content/uploads/2024/06/ECF-87-The-Intercept-Amended-Complaint.pdf" target="_blank" rel="noreferrer noopener">June 21, 2024 – ECF 87 The Intercept Amended Complaint</a>

<a href="https://admin.bakerlaw.com/wp-content/uploads/2024/08/ECF-88-Microsofts-Supplemental-Brief-ISO-Motion-to-Dismiss.pdf" target="_blank" rel="noreferrer noopener">July 8, 2024 – ECF 88 Microsoft’s Supplemental Brief ISO Motion to Dismiss</a>

<a href="https://admin.bakerlaw.com/wp-content/uploads/2024/08/ECF-89-OpenAIs-Supplemental-Brief-ISO-Motion-to-Dismiss.pdf" target="_blank" rel="noreferrer noopener">July 8, 2024 – ECF 89 OpenAI’s Supplemental Brief ISO Motion To Dismiss</a>

<a href="https://admin.bakerlaw.com/wp-content/uploads/2024/08/ECF-90-The-Intercepts-Opposition-to-Defendants-Motions-to-Dismiss.pdf" target="_blank" rel="noreferrer noopener">July 15, 2024 – ECF 90 The Intercept’s Opposition to Defendants’ Renewed Motions to Dismiss</a>

<a href="https://admin.bakerlaw.com/wp-content/uploads/2024/09/ECF-99-OpenAIs-Notice-of-Supplemental-Authority-ISO-Motion-to-Dismiss-The-Intercepts-Complaint.pdf" target="_blank" rel="noreferrer noopener">Aug. 22, 2024 - ECF 99 OpenAI’s Notice of Supplemental Authority ISO Motion to Dismiss The Intercept’s Complaint&nbsp;</a>

<a href="https://admin.bakerlaw.com/wp-content/uploads/2024/09/ECF-91-OpenAIs-Notice-of-Supplemental-Authority-ISO-Motion-to-Dismiss-Raw-Story-Medias-Complaint.pdf" target="_blank" rel="noreferrer noopener">Aug. 22, 2024 – ECF 91 OpenAI’s Notice of Supplemental Authority ISO Motion to Dismiss Raw Story Media’s Complaint</a>&nbsp;

<a href="https://admin.bakerlaw.com/wp-content/uploads/2024/09/ECF-100-The-Intercepts-Notice-of-Supplemental-Authority-ISO-Opposition-to-OpenAIs-Motion-to-Dismiss.pdf" target="_blank" rel="noreferrer noopener">Aug. 27, 2024 – ECF 100 The Intercept’s Notice of Supplemental Authority ISO Opposition to OpenAI’s Motion to Dismiss&nbsp;</a>