PLATFORMS Signal 418
Microsoft exec reportedly called AI scraping the largest theft of labor in human history
↗
Unredacted court filings reveal Microsoft privately called OpenAI's data practices theft while both companies scraped paywalled Times content.
This admission contradicts the companies' fair use defense in the ongoing lawsuit. It highlights the potential economic impact on publishers and the ethical concerns around AI training data. Engineers working on AI models should be aware of the legal and ethical implications of data scraping practices.
Written by elseif from the cluster below · every claim links back to a sourceThe three things worth knowing
Microsoft privately described AI scraping as theft in internal communications.
OpenAI's AI models were said to pose an existential threat to publishers.
The companies allegedly bypassed paywalls and stripped copyright notices from training data.
THE READ
What the cluster adds up to.
The unsealed court filings reveal that Microsoft and OpenAI have been engaged in practices that they privately acknowledged as problematic. A Microsoft executive described the scraping of data for AI training as 'the largest theft of labor in human history,' while OpenAI's leadership acknowledged that their AI models posed an 'existential threat' to publishers. This admission is significant as it directly contradicts the companies' public stance on fair use and the legality of their data practices.
The filings detail how the companies obtained and used content by bypassing paywalls, mass scraping, and stripping copyright notices. This raises serious ethical and legal questions about the methods used to train AI models. The sheer scale of the copying is striking, with OpenAI's mid-training datasets alone containing tens of thousands of copies of works published by various news outlets. This practice not only undermines the economic foundations of publishers but also threatens the employment of the very people who generated the data.
The admissions made in these filings run counter to the fair use defense that AI companies have been relying on. Fair use typically allows for the use of copyrighted work without permission in certain cases, such as parody, news reporting, or criticism. However, the internal communications suggest that the use of this data is not transformative but rather substitutive, directly competing with the original work. This could significantly weaken the companies' legal position in the ongoing lawsuit.
Written by elseif from the cluster below · checked for specifics the sources never containedTHE CLUSTER