# OpenAI and Microsoft foresaw web's 'doom loop' from AI content scraping Internal documents reveal companies anticipated publisher collapse while pursuing AI profits. - **Source:** [The Verge (theverge.com)](https://www.theverge.com) - **Published:** 2026-09-18 - **Topic:** [Data & AI](https://brifing.si/c/data-ai) - **Original article:** - **On Brifing:** https://brifing.si/s/58751eb6-5081-4ebf-8a86-e85e754c13e6 ## The 'doom loop' warning Internal Microsoft documents admitted their AI strategy created a 'doom loop' that would simultaneously degrade model performance and damage the web ecosystem. The company acknowledged this was highly unusual - an end product threatening its own supply chain. Microsoft's director of applied science called it 'the largest theft of labor in human history' while criticizing their fair use claims. ## Publisher existential threat Court filings show OpenAI executives knew ChatGPT posed an 'existential threat' to publishers like NYT. Their policy director admitted creating systems that substitute for cultural labor. Internal documents described ChatGPT as 'the modern newsstand,' with employees noting users have 'no good reason to click' source links after getting AI answers. ## Verbatim copyright violations OpenAI internally acknowledged ChatGPT's tendency to reproduce copyrighted material verbatim. Despite recognizing the importance of preventing memorization to minimize violations, employees admitted GPT-4 'memorized a ton of data' making it 'insanely good at regurgitation.' Examples showed outputs copying full articles from major publishers. ## Search replacement effects Satya Nadella admitted chatbots have replaced traditional search, eliminating the need to visit source sites. OpenAI's experts directly attributed 60% referral traffic drops at publishers to AI summaries. Microsoft conceded almost no content creators intended or were compensated for this use case, calling LLMs 'products that destroy their own supply chain.' ## Corporate contradictions While Microsoft distanced itself from critical internal statements, calling them individual perspectives, the documents reveal systemic awareness. Despite Nadella's claims about licensing paywalled content, an OpenAI rep admitted no efforts were made to detect or remove such content from training data, highlighting policy-practice gaps. ## Economic motivations The filings reveal profit motives behind the AI push, with OpenAI cofounder Greg Brockman focused on 'gazillions' in commercial potential. Microsoft's documents show they proceeded despite recognizing the harm to publishers and their own models, prioritizing financial gains over ecosystem sustainability in what they termed a 'doom loop.' ## Legal strategy exposed The unsealed documents undermine Microsoft's legal position on fair use. Internal admissions that content use violated creator intentions contradict public defenses. Microsoft's attempt to characterize critical employees as 'adversarial' academics contrasts with the documents showing widespread internal concerns about AI's web impacts. ## Industry-wide implications The revelations extend beyond Microsoft and OpenAI, exposing fundamental tensions between AI development and content ecosystems. The documented awareness of harm to publishers, their employees, and ultimately AI model quality suggests systemic risks as companies prioritize short-term gains over sustainable data practices in the generative AI gold rush. ## Sources covering this story - https://www.theverge.com/ai-artificial-intelligence/997633/openai-microsoft-chatgpt-ai-new-york-times-doom-loop-theft-google-zero — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._