Several major news outlets, including The New York Times, asked a judge Thursday to sanction OpenAI for allegedly withholding evidence in a copyright dispute with the ChatGPT maker. The newspapers argue that OpenAI incorrectly claimed it could not search for and produce relevant materials during the discovery process, a move they say has impeded their case.
The copyright lawsuit, which accuses the AI firm of improperly using copyrighted articles to train its large language models, now faces a procedural hurdle over evidence production. If the court grants the sanctions request, it could force OpenAI to hand over additional internal documents, potentially revealing how it selected training data and whether it used paywalled content without permission.
The motion reflects mounting tensions between media companies and AI developers over intellectual property. News organizations have argued that Silicon Valley firms are building billion-dollar businesses on their journalism without compensation, while defendants typically maintain that using publicly available web content for training is protected under fair use.
Public opinion remains divided on the issue. Polling shows that a majority of Americans are unaware of how AI models are trained, but among those who are, more than 60% believe tech companies should pay for copyrighted content used in training, according to recent surveys.
OpenAI has not yet filed a response to the sanctions request. The case could set an important precedent for how copyright law applies to artificial intelligence training datasets, particularly when the data consists of news articles behind paywalls or protected by other access controls.