Microsoft director: AI scraping 'the largest theft of labor in human history'
summary
The New York Times lawsuit against OpenAI and Microsoft reveals internal documents suggesting that AI leadership knew their scraping practices were detrimental to publishers. These documents characterize the use of copyrighted data as an unprecedented theft of labor and an existential threat to the economic foundations of the content supply chain. This admission complicates the companies' legal defense of fair use by showing they were aware of the market damage caused by their models. The case highlights a fundamental conflict between AI development and the intellectual property rights of original content creators.
key points
- Internal Microsoft memos describe AI scraping as the largest theft of labor in human history, acknowledging that creators were neither compensated nor intended for their work to be used this way.
- Data shows that AI tools like Copilot significantly reduced click-through rates for publishers, creating a doom loop where AI models threaten the very sources they rely on for training data.
- OpenAI leadership admitted internally that their chatbot is an existential threat to publishers because it acts as a substitute for original news articles rather than a gateway to them.
who should read it
Tech enthusiasts, legal professionals, and digital publishers who are concerned about the copyright laws and the economic impact of generative AI on journalism.
discussion skeptical
consensus
Commenters generally agree that LLMs engage in a form of intellectual theft by scraping original work and failing to provide attribution, which they view as a corruption of truth. There is a strong sentiment that the lack of provenance in training data is a systemic failure rather than a technical accident. Many argue that the loss of attribution and agency for creators is a significant harm that outweighs the purported gains of generative AI.
top insight
The discussion highlights a 'creditpocalypse' where LLMs not only scrape original terminology but actively hallucinate new authors for those terms when questioned about their origin. This demonstrates that the models are designed to parrot material without any mechanism for recall or sourcing, effectively erasing the original creator from the a digital record.
Generated by an LLM from the article and thread. Check the sources before relying on it.