Low Engagement in Microsoft’s Chatbot for NYT Articles, Reveals Data

In response to copyright claims from publishers, including The New York Times and book authors, Microsoft’s Copilot has stated that it rarely reproduces full sentences from news articles and books. According to new legal filings, the company argues that even substantive chunks that could substitute for the original content are not commonly replicated by Copilot.

As part of the lawsuit’s discovery process, Microsoft provided 8.2 million Copilot chat logs to an expert hired by news publishers. The logs were selected based on keywords that implicated the use of content from the publishers’ websites. Microsoft claims that an analysis of these logs revealed instances where Copilot had common words with news content, with a significant number of matches found. Experts also identified overlaps with works from the Center for Investigative Reporting and authors’ books in the dataset.

The New York Times has refuted Microsoft’s claims, stating that the evidence uncovered during discovery indicates that Microsoft and OpenAI utilized their content to create commercial products that compete with the journalism industry. The lead counsel for The New York Times expressed strong objections to the alleged theft and emphasized the need for accountability from Microsoft and OpenAI. The Center for Investigative Reporting and Authors Guild have yet to comment on the matter.

Microsoft contends that using copyrighted content for AI training datasets should be considered fair use. Despite the reproduction of sections of text by systems like Copilot, Microsoft asserts that the resulting AI models serve different purposes than the original content. The company maintains that the transformative nature of the training process justifies its actions.

The legal battle between publishers, authors, Microsoft, and OpenAI is ongoing, with allegations of product development based on copyrighted works. Despite efforts to streamline the process, the case is moving forward in court. Microsoft has requested a summary judgement, which, if granted, could conclude the case early. The involvement of the Trump administration in supporting OpenAI in the New York Times case adds a layer of complexity to the legal proceedings.