Microsoft says Copilot rarely reproduces NYT articles in chatbot logs
仅事实

Microsoft says Copilot rarely reproduces NYT articles in chatbot logs

摘要

In a court filing, Microsoft contended that its Copilot chatbot seldom copies sentences from New York Times articles or books, citing analysis of millions of chat logs.

Microsoft told a federal court that its Copilot chatbot almost never reproduces full sentences from news articles or books, arguing the AI’s output is largely transformative. The company provided 8.2 million Copilot chat logs to an expert hired by news publishers, selecting those that contained keywords linked to the plaintiffs’ websites. Analysis of that subset showed 59,545 interactions shared at least 16 words with the news content used to train the model, while an independent expert identified 51 instances of "substantial overlap" with material from the Center for Investigative Reporting. Another expert found only 24 responses containing 30 or more matching words, and just 10 of 212 books examined showed any matches.

"The documents and testimony uncovered during discovery lead to only one conclusion: Microsoft and OpenAI stole from The New York Times to make commercial products that substitute for its journalism, threaten its business, and undermine its industry," said Ian Crosby, lead counsel for The New York Times.

Microsoft argued that the low incidence of matching text supports a fair-use defense, emphasizing that the AI’s purpose differs from the original works and that occasional excerpts do not negate the system’s transformative nature. The filing seeks a summary judgment that would end the consolidated lawsuit brought by publishers and authors, who allege that Microsoft and OpenAI built products that directly compete with their copyrighted material. The case remains pending before a judge.

继续阅读——免费

创建免费账户以关注新闻。无需银行卡。