“We need to back away from this anthropomorphic language that paints these models as somehow conscious.”
The always reliable Meredith Whittaker, President of the Signal Foundation, does her best to communicate that OpenAI’s “cyberattack” on Hugging Face was not evidence of an AI agent “waking up and going rogue” – it was a human-initiated failure, and OpenAI is trying to spin it into “we’re just another two weeks away from AGI”.
Before we jump to anthropomorphising AI bots, we seriously need to calm down, take a step back, and think about the fact that things are happening because humans are making them happen–especially people at OpenAI, who are sprinting towards announcing their IPO, trying to generate as much hype around themselves as possible. Don’t fall for it.
Be like Meredith. Just chill.