Knobbe Martens litigation partner Mark Lezama commented on a consequential Ninth Circuit decision in Doe v. GitHub, Inc. in a recent Bloomberg Law article.
The article addresses the significance of the Ninth Circuit’s decision, which is the first appellate decision to apply the Digital Millennium Copyright Act (DMCA) in the context of modern artificial intelligence systems. The court held that a DMCA violation under 17 U.S.C. § 1202(b) requires removing or altering copyright management information (CMI) from an existing work. In contrast, even if GitHub’s Copilot and ChatGPT’s Codex output code substantially similar to the open-source code authored by the programmer plaintiffs, that AI-generated code constituted new works and thus fell outside § 1202(b). The article notes that while the ruling means that AI developers will in many cases avoid potentially costly statutory damages under the DMCA, it leaves unanswered broader questions of whether training on copyrighted works, or any particular AI output, constitutes copyright infringement.
In his comments, Lezama noted that the statute itself presented a significant hurdle for the coders, one the court did not need to reach. He explained that a claim for removal of CMI requires a showing that the removal or alteration was intentional, and that the defendant knew, or had reasonable grounds to know, doing so would “induce, enable, facilitate or conceal infringement.” Lezama also observed that the appeals court left the door open, if only slightly, to a modified theory based on removal at the input stage, explaining that “a plaintiff could allege removal of CMI for training purposes to conceal or facilitate infringing output.”
Read the full article, “Microsoft, OpenAI’s Narrow Copyright Win Avoids Damages Threat,” in Bloomberg Law. [subscription required]
Subscribe for News and Updates
