On forums and social media, language models have become objects of reverence. Users quote ChatGPT as if reciting holy texts.

I call this phenomenon “The Cult of the Stochastic Parrot.”

Origins of an Odd Name Link to heading

Emily Bender and her colleagues coined “stochastic parrot” in a 2021 paper. It captures what large language models do: play a complex game of “predict the next word” without understanding anything they’re saying. Like a parrot that sounds convincing but makes no sense.

The “cult” part comes from watching forums. “ChatGPT told me…” appears in every other post, each proclamation treated as gospel. The fervor spreads unchecked.

The name also borrows from Slack’s Party Parrot emoji culture; a bouncing bird helps lighten up a serious topic.

Our Brain’s Wiring Link to heading

Pattern recognition shapes how we see the world, from faces in clouds to personalities in machines. This cognitive quirk now leads people to attribute consciousness to sophisticated mimicry. The certainty with which users defend sentience would be remarkable if it weren’t so widespread.

Mistaking tools for something smarter than they are has consequences. Blind trust in outputs and missed flaws in oversight create real problems while more urgent threats develop unnoticed.

The Moonwalking Bear Has Entered the Chat Link to heading

The online world obsesses over ChatGPT’s supposed feelings. Meanwhile, facial recognition systems lead to wrongful arrests, and synthetic media spreads unchecked across social platforms. The Cult of the Stochastic Parrot and our hand-wringing over digital theology blinds us to actual problems.

Back to Reality Link to heading

Pattern matching tools need verification, not worship. Like counting basketball passes while missing the moonwalking bear, our fascination with AI consciousness distracts us from the real show. The future of AI depends less on its imagined sentience and more on how we choose to use it.