Research
Courts have now ruled AI training on copyrighted books to be "spectacularly transformative" fair use in two major U.S. decisions, but the NYT-led coalition's sanctions motion — alleging OpenAI destroyed evidence of how journalism was used in training — could reshape that picture if spoliation is proven, since the core evidentiary problem is that even OpenAI's engineers cannot reliably trace which content drove which model behaviors. OpenAI is simultaneously licensing content from AP, Axel Springer, the Financial Times, News Corp, and Le Monde while contesting the NYT lawsuit, revealing a dual strategy that raises an unresolved distributional question: if training is lawful fair use, why pay some publishers and not others? The deepest tension the briefing surfaces is not doctrinal but structural — whether the right remedy for journalism's compounding economic crisis is copyright litigation, collective licensing, statutory levies, or equity deals remains genuinely contested, and the experiences of small and mid-size publishers, who make up most of the press ecosystem, are almost entirely absent from the evidence base.
The lawsuit alleges that OpenAI trained its artificial intelligence (AI) models on NYT content without authorization and that OpenAI and Microsoft's products ...
"This motion asks the court to punish OpenAI for hiding and destroying evidence showing how ChatGPT was trained on stolen journalism," said ...
Anthropic and Meta decisions both suggest fair use can apply to AI training but they diverged on two key issues: the impact of using pirated ...
The arrangement sees OpenAI licensing part of AP's text archive, while AP will leverage OpenAI's technology and product expertise.
OpenAI said the deal covers “select attributed summaries, quotes and links to FT journalism.” It didn't specify when the content will become ...
Axel Springer and OpenAI have announced a global partnership to strengthen independent journalism in the age of artificial intelligence (AI).
Part 2 was published on January 29, 2025, and addresses the copyrightability of outputs created using generative AI. Guidance for Works Containing AI-Generated ...
The New York Times sued OpenAI in December 2023 for copyright infringement, accusing the AI creator of using its article archive to train ...
On June 23, 2025, the United States District Court for the Northern District of California issued a significant order in Bartz, et al.
The courts held that, on the particular facts of those cases, training language models on copyrighted books was transformative and fair use, ...
News Corp and OpenAI to Partner in Supporting the Highest Journalistic Standards. News Corp will share journalistic expertise to help ensure ...
Le Monde has just signed a multi-year agreement with OpenAI, its ChatGPT tool. It covers both the training of artificial intelligence models ...
The New York Times is accusing OpenAI of deliberately hiding evidence and logs that could link copyrighted content to ChatGPT's training data.
Stay up to speed on the rapid advancement of AI technology and the benefits it offers to humanity.
The AI Act entered into force on 1 August 2024 and became applicable on 2 August 2026, with some exceptions: prohibited AI practices and AI literacy obligations ...
The consultation included proposals to establish an exception to copyright law for text and data mining, including for commercial activities.
Japan has provided a blanket copyright exception for AI training is miles off-base. had opened the taps to permissionless copying of ...
The report misapplies long-settled fair use principles and ultimately puts a thumb on the scale in favor of copyright owners at the expense of ...
by D Inserra · Cited by 2 — The most significant threats to the expressive power of AI are government mandates and restrictions on innovation. Cato Institute Policy ...
Sign up to read the full research briefing
Sign up