Editors’ Summary: This post includes a podcast episode and an accompanying essay summarizing the conversation. The conversation is focused on the way the media reports about AI and its own so-called agency. The recent headlines about an OpenAI model “hacking” another AI company, Hugging Face, is an example of the rhetorical strategy. The author notes that in that instance OpenAI developers intentionally removed that model’s cybersecurity blocks and tested it with a live network connection. In other words, “they trained a model to find exploits and then acted surprised that it found them.” The authors describe this phenomenon as “the system from nowhere,” which implies AI without people. This tendency obscures the human creators and designers of AI, in favor of the myth of a self-improving, self-aware system. They point to how this absence of the human in AI shapes policies, and often displaces the people involved in less powerful positions, such as data cleaners or people living near data centers.
Editors’ Choice: The System From Nowhere
