Unsealed NYT v. OpenAI filings reveal Microsoft execs calling AI training 'largest labor theft in human history'

On September 18, a court unsealed the summary judgment motions in the NYT v. OpenAI/Microsoft copyright case, revealing previously redacted internal statements—many of the most striking phrases coming from the AI companies' own employees and executives. Commentators including Gary Marcus and Jason Kint view these as rare 'admissions' and an explanation for the heavy redactions. The case remains at the summary judgment stage with no legal findings, but the disclosures have materially shaken the broader AI training-data copyright debate.

Confirmed

• Per reporting by Tom's Hardware, Ars Technica and NYT reporters Scott Nover and Gerrit De Vynck, Microsoft Director of Applied Science Brent Hecht described LLM training-data scraping in a January 2023 internal memo as 'an unprecedented, shocking theft' and possibly 'the largest labor theft in human history.' A 2023 internal Microsoft document also shows employees debating whether OpenAI's practices constituted the greatest labor theft ever and worrying they threatened model quality.
• Per WSJ, OpenAI and Microsoft executives internally acknowledged their AI outputs were 'increasingly substitutive' and posed an 'existential threat' to publishers—showing leadership long knew scraping news to train models could harm publishers.
• Plaintiffs allege OpenAI used a 'paywall hack' to bypass subscriptions and scrape NYT content and removed copyright notices (m7, m14, m15, m16): after an employee described the hack, OpenAI co-founder and president Greg Brockman replied 'ah nice' (m5–m12, m9); media industry figure Jason Kint called it 'infuriating' and a stain on both OpenAI and Brockman.
• Per emmanuelvivier, unsealed filings show Microsoft internally anticipated AI scraping and generated content could drag the internet into a 'doom loop' killing 'the entire Web' (m1, m13, m14).

Unconfirmed

• Whether Brockman's 'ah nice' reply actually constitutes legal liability such as a CFAA violation: commentators including matthewstoller have raised the question, but no legal determination exists yet.

Why it matters

• These statements come from the defendants' own internal documents and testimony—rare 'admission'-style evidence that could shape the case and the wider training-data copyright fight. Gary Marcus and others argue these phrases hand plaintiffs powerful arguments; Jason Kint quipped the internal statements wrote the NYT's lede for it.

2026-09-18 ~ 2026-09-20 · 22 related posts

• NYT accuses OpenAI of deceptive paywall circumvention, citing Brockman's 'ah nice' reply — Polymarket · 2026-09-18
• Polymarket prices 15% odds OpenAI announces AGI by 2026, as NYT flags paywall hacking — Polymarket · 2026-09-18
• Unredacted filings: Microsoft scientist called LLM training "an astonishing theft of unprecedented proportions" — GarrisonLovely · 2026-09-18
• Unsealed filings: Microsoft exec called AI scraping 'largest theft of labor in human history' — RebeccaBellan · 2026-09-18
• Court unseals motions in NYT v. OpenAI, exposing 'largest theft of labor' internal language — kyliebytes · 2026-09-18
• [source] Court Docs Show Greg Brockman Responded "ah nice" to NYT Paywall Scraping Hack — kyliebytes · 2026-09-18
• [source] Microsoft exec called AI training 'largest theft of labor in human history', unsealed filing shows — ScottNover · 2026-09-18
• OpenAI president's "ah nice" reply to NYT paywall hack sparks CFAA violation accusations — GaryMarcus · 2026-09-18
• OpenAI and Microsoft execs knew scraping news sites could harm publishers, filing says — TuhinChakr · 2026-09-18
• Unsealed NYT v. OpenAI Filings: Execs Admit Products Are 'Largely Substitutive' — GaryMarcus · 2026-09-18
• Court docs reveal Microsoft internally predicted a web "doom loop" from AI — emmanuelvivier · 2026-09-19
• Microsoft exec calls LLMs 'largest theft of labor in history' as internal doc warns of a web 'doom loop' — CackleRooster · 2026-09-19

10 near-duplicate retellings: nordicinst · GaryMarcus · AlexTensor · GaryMarcus · emmanuelvivier · StefanoGogioso · MarietjeSchaake · eyishazyer · emmanuelvivier · beingmodest