Microsoft executive called AI scraping "the largest theft of labor" in unsealed filings

SEATTLE — Newly unsealed court filings made public on Sept. 17, 2026, reveal that a Microsoft executive privately characterized OpenAI's data scraping practices as "the largest theft of labor in human history," even as both companies actively scraped paywalled content from The New York Times to build foundational training datasets.

Strategic Context

According to documents highlighted by TechCrunch on Sept. 17, 2026, internal communications show that major developers were fully aware of the friction surrounding web scraping. The unredacted records detail how the firms ingested paywalled journalism to train large language models while circulating internal warnings that these data-harvesting operations would severely undermine traditional news publishers.

Financial & Macro Implications

The disclosures expose the operational reality between platform providers and content creators. By relying on uncompensated web scraping to secure text corpora for model training, the firms avoided licensing costs that would otherwise impact operating expenses. The unsealed documents demonstrate that leadership inside the ecosystem recognized the threat this practice posed to publishers, framing the mass acquisition of proprietary text as a direct appropriation of intellectual labor.

Forward Outlook

As these unsealed court records enter the public record, operators and allocators must monitor how judicial rulings treat unauthorized data ingestion for commercial AI training. The revelations threaten to reshape negotiations over data licensing fees and increase compliance overhead for platform operators scaling next-generation models.

Source: TechCrunch: Microsoft exec called AI scraping ‘the largest theft of labor in human history,’ new unredacted filings reveal.