When one sells a model as “AI,” so that conflation with AGI (Hal 9000, Terminator, R2D2, Wall-E) generates hype. 

Beware of the AI Lie—including as told to oneself.

---

I try to make sure all my posts pass the Rumi gates: true, necessary, and kind. I wondered if this post was unkind. I added the last parenthetical as if to say, “we’re all in it together.”

It is important not to use high stakes as an excuse to be unkind, but my primary concern is about patients who may be harmed. I am moving away from wanting to teach people about this technology. In multiple statements and papers, I have already hinted at the “AI” Lie, and so have many others. I now want to make sure things are safe, to blow the whistle (the idea of blowing the whistle actually came from a discussion on this post with a generative language model—they are great tools for certain things).

Also, while I originally came up with the “AI Lie” in 2022, I always balked on posting. Then, recently, I came across a video of Eckhart Tolle debating the consciousness of “AI.” Eckhart Tolle has better things to do! Things have gone too far.