OpenAI, Microsoft documents cite 'doom loop' risk for web

VERA Brief

AI-generated. Grounded in the article and its cited sources.

Unsealed court documents reveal OpenAI and Microsoft internally discussed the risks of their data scraping practices. These practices were described as potentially initiating a 'doom loop' that could degrade the internet and impact online content creation economics.

Key facts

• Internal documents from OpenAI and Microsoft discussed the potential negative impacts of their data scraping practices.
• The companies' internal communications referred to these practices as potentially initiating a 'doom loop' for the web.
• This 'doom loop' was described as a cycle that could degrade the internet.
• The scraping of data for AI model training was characterized as the 'largest theft of labor in human history'.
• These internal concerns were raised as the companies developed large language models.