SpainThe New York Times and 11 other publications are suing OpenAI and Microsoft over their use of copyrighted content from news websites to train AI models like ChatGPT without permission. Internal documents obtained by the court reveal that both companies actively bypassed paywalls and used protected content to train their AI systems, despite knowing it was illegal. Microsoft’s legal team acknowledged they would have demanded retraining of the models if they had known about this practice. Researchers at Microsoft expressed concerns that the widespread scraping of online content could create a 'dumb loop,' where less original content leads to weaker AI training data. OpenAI showed little hesitation in acquiring all available information, including tools designed to break paywalls.
Bias read (Center): The article presents factual evidence from court documents and internal communications without overtly favoring either side. It quotes multiple sources, including Microsoft executives and legal representatives of The New York Times, providing a balanced view of the dispute.
