In re OpenAI, Inc. Copyright Infringement Litigation
The ongoing legal battle against OpenAI revolves around its alleged use of Ziff Davis’ online publications in the development of its GPT series of large language models. This litigation has seen a district court ruling that denies OpenAI’s motion to dismiss multiple claims, including contributory copyright infringement, removal of copyright management information, and distribution of works without that information. However, the court has granted motions to dismiss claims related to the circumvention of technological measures based on the disregard of robots.txt instructions and has partially dismissed trademark dilution claims.
Background of the Case
Ziff Davis, a prominent digital media entity, alleges that OpenAI unlawfully utilized its copyrighted works in the training of its large language models like ChatGPT. The claims encompass a range of legal issues, including:
- Direct copyright infringement due to the use of Ziff Davis’ works without permission.
- Contributory copyright infringement.
- Violations under the Digital Millennium Copyright Act (DMCA), including circumvention of technological measures and removing copyright management information.
- Federal trademark dilution and state law claims for common law unjust enrichment.
Court Rulings
The court’s analysis led to several significant rulings:
1. Unjust Enrichment Claims
The district court dismissed Ziff Davis’ claim for common law unjust enrichment, agreeing with OpenAI that this claim was preempted by Section 301 of the Copyright Act. The ruling established that the content in question falls under the Copyright Act’s jurisdiction, meaning state law claims cannot seek to protect rights equivalent to those protected under federal law.
2. Circumvention of Technological Measures
Ziff Davis argued that its robots.txt files were a technological measure that OpenAI circumvented. However, the court determined that these files do not constitute effective technological controls under the DMCA. The court ruled that merely ignoring these instructions does not equate to circumvention.
3. Distribution of Works with CMI Removed
OpenAI contended that Ziff Davis did not adequately allege that it distributed complete copies of its work without copyright management information. The court, however, found that Ziff Davis’ claims that users could prompt ChatGPT to generate identical versions of its copyrighted works were enough to support allegations of distribution at this stage of litigation.
4. Trademark Dilution Claims
The court partially upheld Ziff Davis’ claims regarding trademark dilution, finding that the mark associated with its platform, Mashable, is sufficiently famous. However, the court dismissed claims related to other brands like CNET and PCMag due to insufficient evidence of their fame.
5. Contributory Copyright Infringement and Removal of CMI
The court denied OpenAI’s motion to dismiss claims related to contributory copyright infringement and the removal of copyright management information. The court found that Ziff Davis had plausibly alleged that OpenAI had knowledge of infringement occurring through its models.
Stays of Proceedings
OpenAI sought a stay of proceedings on specific claims, but the court found that doing so would not significantly aid in expediting the complex litigation. Thus, the stay was granted regarding certain models not directly implicated in the current litigation.
This case exemplifies the evolving intersection of technology and copyright law, particularly as it pertains to the capabilities and responsibilities of AI systems in handling copyrighted content.
