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

Documents unsealed in a lawsuit against OpenAI and Microsoft reveal top executives privately describing their AI systems as the product of unprecedented theft from creators and publishers. The filings come from a case brought by Mother Jones and other outlets, which accuse the companies of training models on copyrighted material without permission or compensation.

Microsoft’s own director of applied science, Brent Hecht, warned internally that AI development has created a “doom loop” that will “threaten the economic foundations of its essential suppliers.” He noted the process would degrade the quality of information available online as original reporting and content creation become unsustainable. These statements emerged not from critics but from within the company pushing the technology hardest.

The companies have framed their work as essential progress that could reshape society. At the same time, executives have acknowledged the scale of what they took. One description in the court papers calls it “the largest theft of labor in human history.” The material includes reporting from newsrooms that invested years building audiences and trust, only to see it scraped and repurposed into chatbots that now compete directly with those same outlets.

This is not an abstract debate over innovation. Real reporters, photographers, and independent writers produced the content that fed the models. Their labor now powers products valued in the hundreds of billions, while the original creators face shrinking revenues and uncertain futures. The pattern mirrors other moves by concentrated tech power, where rules are treated as optional when they stand in the way of rapid growth.

The companies continue to argue they should face few limits even as they flag existential risks from the same systems. Public statements emphasize safety and responsibility. Internal documents show a different calculation: proceed quickly, absorb whatever material is available, and deal with consequences later. That approach has already produced tools that can mimic writing styles and summarize reporting without attribution or payment.

Lawsuits from multiple publishers are testing whether copyright protections still apply when the end product is an AI model rather than a direct copy. The early evidence suggests the firms understood the legal exposure but calculated the rewards were worth the risk. Ordinary Americans who rely on independent journalism for information about government and business now face an industry under pressure from the very technology built on its output.

The gap between the companies’ public posture and their private assessments continues to widen. One side talks about benefiting humanity. The other side counts the cost to the people whose work made the systems possible in the first place.