Gracenote Sues OpenAI Over Alleged Copyright Misuse in AI Training Data

Gracenote’s complaint alleges that OpenAI scraped the company’s proprietary metadata—information that uniquely identifies songs, movies, TV shows and other media—for use in training its large‑language models, without securing a licensing agreement. The filing, which was lodged in Manhattan federal court on Tuesday, says OpenAI “systematically ignored” Gracenote’s repeated outreach to negotiate a data‑licensing deal, according to the Reuters legal‑litigation feed. Gracenote, a Nielsen subsidiary that powers the “song‑identification” features in millions of consumer devices, claims its copyrighted databases were incorporated into the training corpus that underpins ChatGPT and related products, a violation that could expose OpenAI to statutory damages under U.S. copyright law.

The suit also seeks a preliminary injunction to halt OpenAI’s continued use of Gracenote’s data, demanding that the AI firm preserve all copies of the allegedly infringing material and provide an accounting of the models that have been trained on it. In its complaint, Gracenote points to internal OpenAI communications that, per the filing, reveal awareness of the data’s proprietary nature and a decision to proceed regardless. The company argues that the “unfair competition” resulting from the unlicensed use of its metadata undermines the value of its licensing business, which fuels revenue streams for content owners and distributors across the entertainment ecosystem.

OpenAI’s legal team has not yet responded publicly, but the filing notes that the company has previously defended similar claims by arguing that its training data is derived from publicly available sources and that any copyrighted material is transformed under the doctrine of fair use. Legal analysts cited by Engadget have highlighted that the “fair use” defense is increasingly tenuous in the context of massive, non‑human‑curated datasets, especially when the plaintiff can demonstrate that the data was harvested in bulk and used to improve a commercial product. Gracenote’s case, therefore, may become a bellwether for how courts assess the balance between AI innovation and the protection of curated, copyrighted databases.

If the court grants Gracenote’s request for injunctive relief, OpenAI could be forced to purge a substantial portion of its training corpus, a move that would have ripple effects across its suite of offerings. The company’s API customers—including enterprises that rely on ChatGPT for content generation, code assistance and data analysis—might see degraded performance or delayed feature rollouts while OpenAI rebuilds its models with alternative data sources. Industry observers have warned that such a setback could accelerate the push for more transparent data‑licensing frameworks, a trend already evident in recent negotiations between AI firms and music publishers.

The lawsuit arrives amid a wave of litigation targeting AI developers for alleged data misuse, with Gracenote joining other metadata and content providers that have taken legal action this year. As the case proceeds, it will test the limits of existing copyright statutes in the era of generative AI and could shape the contractual norms that dictate how proprietary datasets are accessed, licensed and protected. For now, the filing underscores the growing friction between the rapid expansion of AI capabilities and the entrenched interests of content owners seeking to safeguard their intellectual property.