In a recent episode of the 404 Media podcast, the hosts discussed court records in which OpenAI and Microsoft reportedly admit that their large language models are damaging the internet and that they have stolen intellectual property on an unprecedented scale. These are not casual remarks but formal admissions from legal filings, suggesting the companies themselves acknowledge the harmful side effects of their technology.
The same episode also covered a practical demonstration of AI misuse: a journalist managed to take over a real band's Spotify page using AI-generated music, finding the process disturbingly easy. This highlights how AI can be weaponized to impersonate artists and flood platforms with synthetic content, undermining trust in creative works.
Additionally, the podcast examined an AI agent platform called iLands, which the hosts describe as basically spam. The platform appears to automate the generation and distribution of unsolicited messages, contributing to the growing problem of AI-driven inbox flooding. Together, these stories paint a picture of an AI ecosystem that, by its own admission and through real-world examples, is eroding the quality and authenticity of online spaces.