Every time I express that I think it's shitty to say that genAI "hallucinates," people get weirdly pissy and combative with me.
But until real human beings who experience hallucinations are treated like human beings worthy of respect and dignity...I think perhaps it's shitty to keep associating hallucinations with being non-human, dangerous, and a liar.
It actually really fucking sucks that hallucinating is seen as monster-movie-scary, dangerous, inhuman, and worthy of scorn and/or mockery. It actually really fucking sucks that people who hallucinate or experience delusions can't get compassionate help or understanding, because everyone thinks it's an automatic sign of being dangerous, violent, and a liar. It's a massive problem that people look at hallucinations as "dangerous crazy person shit" that should be gotten rid of or locked away.
I find it crass, at best, to flippantly use the term "AI hallucinations" when real-life people get abused, imprisoned, and killed because they experience hallucinations.
But what do I know, I'm just a crazy person. ¯\_(ツ)_/¯
People in the notes keep proposing alternatives or asking me what they should say instead, and I really think y'all are overthinking it. You don't need to coin an entire new term. You can just say "AI generates misinformation." If you need it as a noun, "AI misinformation."
If you want to be provocative, "AI spits out garbled bullshit" works, too.
My spouse is a computer programmer. 'AI hallucination' is the real term that you are meant to use. It doesn't matter if people like it or not. It matters that it's factual and correct.






















