S&P 500, Nasdaq Hit Records on AI Stocks and Falling Oil

Internal documents from a copyright lawsuit against OpenAI and Microsoft have exposed how top executives at these companies viewed their training of large language models as an unprecedented appropriation of others' work. The revelations come from court filings in a case brought by Mother Jones and other publishers, who argue that the firms built their valuable AI systems by ingesting copyrighted articles without permission or compensation.

Microsoft's director of applied science, Brent Hecht, described the process in stark terms, warning of a "doom loop" where AI systems undermine the economic foundations of news outlets and other content creators that supply the internet's information. Hecht's comments, uncovered through the litigation, align with the plaintiffs' characterization of the companies' actions as perhaps "the largest theft of labor in human history." The brief filed on September 4 details how OpenAI and Microsoft scraped vast troves of published material to power ChatGPT and related tools, bypassing traditional licensing agreements that have long sustained journalism.

This approach allowed the companies to amass enormous value while contributing little back to the ecosystem that generated the data. Similar suits from The New York Times and The Intercept underscore the pattern, with publishers contending that the AI boom rests on systematic disregard for intellectual property rights. Executives at the firms have at times acknowledged the risks internally, even as they publicly push for minimal regulatory hurdles and continued rapid development despite warnings of broader societal harms.

The documents reveal a pattern of calculated risk-taking. Rather than negotiate with content creators or invest in sustainable data partnerships, the companies proceeded with training methods that critics say amount to wholesale extraction. This has left newsrooms and independent journalists facing an uncertain future, as AI-generated summaries and chatbots threaten to siphon audiences and advertising revenue without replenishing the original reporting.

Legal experts following the cases note that the internal admissions could strengthen arguments for fair compensation or restrictions on unauthorized use of protected material. The publishers seek not only damages but also changes to how future models are built, insisting that labor and creativity deserve protection even in an era of automated systems. For now, the filings provide a rare unfiltered look at how Silicon Valley weighed the benefits of speed against the rights of those whose work enabled the technology in the first place.