MS Internal Docs Admit LLMs Destroyed Web via 'Doom Loop'

AI executives at Microsoft and OpenAI have acknowledged that large language models (LLMs) are predatory technologies, as critics have long pointed out. A Microsoft executive described this as 'shocking theft on an unprecedented scale' and 'the greatest theft of labor in human history.' Microsoft internal documents warned that generative AI products are creating a 'doom loop' that is killing 'the entire web.' The documents noted that millions of people will come to view the 'sucking in' of their work by large models as theft, pointing out that content creators did not consent to this usage and received no compensation.