Microsoft Claims Minimal Copyright Infringement by Copilot in NYT Lawsuit Filings

Microsoft's Defense Against Copyright Claims

Microsoft has filed legal documents in response to copyright infringement lawsuits brought by The New York Times and various authors. The company contends that its Copilot chatbot rarely reproduces significant portions of copyrighted content, such as news articles or books, that could substitute for the originals.

Analysis of Chat Logs

As part of the lawsuit's discovery process, Microsoft provided 8.2 million Copilot chat logs to an expert representing news publishers. These logs were specifically selected for containing keywords related to the plaintiffs' websites. Microsoft's analysis indicates that fewer than 1% of these logs, specifically 59,545, contained at least 16 words in common with news content used to train the AI model. Similarly, an expert in the authors' suit found only 24 responses with at least 30 matching words from books within the 8.2 million conversations.

The New York Times' Disagreement

The New York Times has rejected Microsoft's conclusions. Ian Crosby, lead counsel for The Times, stated that the evidence uncovered during discovery points to Microsoft and OpenAI having used The New York Times' content to create commercial products that compete with its journalism. The Times seeks accountability for what it describes as theft.

Fair Use Argument

Microsoft's argument centers on the concept of fair use, asserting that using copyrighted material for AI training datasets should fall under this legal doctrine. The company maintains that while systems like Copilot utilize copyrighted content, the resulting AI systems serve significantly different purposes than the original works. Microsoft concludes that occasional reproduction of text sections does not undermine the transformative nature of large language model training.