The New York Times alleges OpenAI used deceptive and even transgressive methods to bypass paywalls, including an exchange suggesting approval from Brockman for a paywall-bypassing technique.

New York Times court filings claim OpenAI copied millions of copyrighted articles without permission to train its AI models and employed deceptive means to access paywalled content.

Microsoft CEO Satya Nadella testified that he would have required OpenAI to retrain if paywalled data had been scraped, though Microsoft said his remarks referred to broader information-access changes rather than core copyright issues.

The case fits into a broader wave of AI copyright lawsuits that could reshape how AI firms source and use copyrighted material for training.

OpenAI argues that training on publicly available material constitutes fair use, likening its approach to how search engines operate, emphasizing transformations and claiming its products are not substitutes for publishers.

The filing suggests OpenAI acted for commercial gain and could threaten publishers, referencing a 2017 remark by OpenAI co-founder Greg Brockman.

A Microsoft director is described in the filing as calling using copyrighted material to train large language models a form of theft, though Microsoft later clarified the remarks reflected an employee’s view rather than a legal position.

OpenAI has rejected the NYT’s claims, arguing the lawsuit seeks an excessive payout and could hinder technological progress.