OpenAI and Microsoft knew they were starting a ‘doom loop’ for the web

According to a report by The Verge, newly unsealed court documents from the New York Times' ongoing lawsuit against OpenAI and Microsoft have provided a rare look into the internal discussions of the tech giants regarding their data training practices. The filings suggest that employees at the companies were aware of the potential long-term consequences their data scraping methods could have on the broader web ecosystem, with some internal communications reportedly referencing a potential 'doom loop' that could degrade the quality of online information.

The documents, which were made public as part of the discovery phase of the litigation, highlight the tension between the rapid development of large language models and the sustainability of the digital content sources they rely upon. The New York Times has argued that the unauthorized use of its intellectual property constitutes a significant infringement, while the defendants have maintained that their activities fall under the scope of fair use and are essential for the advancement of artificial intelligence.

This development marks a significant escalation in the legal battle, which is being closely watched by media organizations and tech companies alike. The outcome of this case is expected to set a major legal precedent for how AI developers can legally source training data from copyrighted material. Neither OpenAI nor Microsoft has provided a detailed public rebuttal to the specific internal characterizations cited in the report, though both companies continue to defend their development processes as industry-standard and legally compliant.