# Microsoft scientist calls AI training the largest theft of labor

Digest AI · Policy & Regulation · published 2026-09-21T07:24:00Z · updated 2026-09-21T16:08:00Z

Canonical: https://digestai.news/story/microsoft-scientist-calls-ai-training-the-largest-theft-of-labor

## Summary

Dr. Brent Hecht, Microsoft’s Director of Applied Science, described AI model training as “the largest theft of labor in human history” in a filing for the consolidated OpenAI copyright lawsuit before Judge Sidney H. Stein. Hecht’s comments are presented as his individual view, not a legal analysis, and Microsoft later said the remarks reflect only one employee’s perspective.

The plaintiffs – including The New York Times, the Daily News Group, the Center for Investigative Reporting and Ziff Davis – cite internal Microsoft data showing click‑through rates dropping dramatically when users see Copilot results versus traditional Bing: 87%‑93% for the Times sites, 83%‑91% for Daily News, and 51%‑94% for Ziff Davis. The brief also lists the volume of copyrighted material in OpenAI’s training sets, noting more than 91,692 copies of plaintiffs’ works overall, with WebText2 containing at least 6,552 Times articles, 18,609 Daily News pieces and 66,780 Ziff Davis items, plus a Mango dataset of at least 160,903 unique plaintiff works and a New York Times Annotated Corpus of over 1.8 million articles.

The United States filed a statement of interest arguing that training on copyrighted text is fair use, while OpenAI calls the lawsuit an “undeserved payday.” The case could determine whether large‑scale data scraping for foundation models constitutes copyright infringement, shaping future AI development and publisher rights.

## Key points

- Brent Hecht called AI training “the largest theft of labor” in the OpenAI copyright case filing.
- Click‑through rates fell 87‑93% (Times), 83‑91% (Daily News) and 51‑94% (Ziff Davis) when users saw Copilot vs Bing.
- OpenAI’s training data include over 91,692 copies of plaintiffs’ works, with 6,552‑66,780 copies per source and 160,903 in the Mango set.

## Why it matters

The allegation frames AI training as massive copyright theft, and the court’s ruling could set precedent for how future AI models use copyrighted content, affecting publishers and AI developers alike.

## Sources

1. [A Microsoft scientist called AI training the largest theft of labor](https://thenextweb.com/news/microsoft-theft-of-labor-doom-loop-unsealed-brief-openai) (thenextweb.com, 2026-09-21)
2. [Historic lawsuit sparks flurry of option activity in NY Times stock](https://cnbc.com/2026/09/21/historic-lawsuit-sparks-flurry-of-option-activity-in-ny-times-stock.html) (CNBC Technology, 2026-09-21)

Part of the developing story: [Media Giants Sue AI Over Labor Theft](https://digestai.news/thread/microsoft-exec-called-ai-scraping-largest-theft-of-labor-in-human-history) (4 stories)

## Cite

Digest AI, "Microsoft scientist calls AI training the largest theft of labor", 21 September 2026, https://digestai.news/story/microsoft-scientist-calls-ai-training-the-largest-theft-of-labor

---

Written by Digest AI's editorial model from the linked sources; the sources are the record. Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse
JSON: https://digestai.news/story/microsoft-scientist-calls-ai-training-the-largest-theft-of-labor.json