AI training built on fair use looks shaky when the companies' own people call it "astonishing theft"

Newly surfaced internal communications and sworn testimony are challenging the legal foundation that OpenAI and Microsoft have used to justify training AI models on copyrighted works without compensation.

According to The Decoder, a Microsoft director described the practice as the largest theft of labor in human history in internal correspondence. Separately, OpenAI's head of ChatGPT reportedly acknowledged that the company's products are largely substitutive, directly contradicting the fair use argument that AI outputs transform rather than replace original works.

These statements are significant because fair use doctrine hinges on whether a use is transformative and non-substitutive. If courts give weight to admissions from the companies' own leadership, it could weaken a central legal defense in ongoing copyright litigation brought by authors, publishers, and other rights holders.

The disclosures could also pressure AI developers to secure more explicit licensing agreements for training data, potentially raising costs and complicating the data pipeline for future model development across the industry.

This page provides an editorial summary based on publicly available information. It is not a republished article. Use the source link below for the original report.