Harsh truths to save you from ChatGPT psychosis

Please pass this on to that one family member who needs it

I am currently dealing with multiple people in my life who seem deep in the throes of what Futurism calls “ChatGPT psychosis.” I suspect you may know people like this as well. It is alarming and perplexing.

People have been confusing chatbots for sentient conversation partners since the 1960s, but what’s going on here feels new. Sometimes the victims of ChatGPT (or Claude, or Gemini) psychosis come to believe that AI is a godlike entity—but it never starts there. First AI makes them feel that they themselves are superintelligent. That their most sophomoric ideas are brilliant, that their vague prompts qualify as original thinking, or that their uninformed speculations are important contributions to scholarly fields they have never studied.

Once you come to believe that you are sort of a minor cybernetic deity, you have lost touch with reality in a subtle, terrifying way. You are at the mercy of whatever weird fantasy the LLM spits out next.

The major LLMs are all in a sycophantic phase right now, which doesn’t help, but I doubt that “make the chatbots less encouraging” is an easy fix here. We are dealing with a technology uniquely suited to snipe intelligent, well-educated people into believing they are much, much smarter than they really are. That is an addictive sensation, not easily quit.

That means you, yes you, reading this, are at risk.

If you talk to LLMs at all—and at this point, who doesn’t?—it might not be a bad idea to post these reminders next to your computer, just to protect your own mind against infohazards:

1. AI will not make you smarter. It will make you faster at retrieving (possibly correct) answers to certain questions. It will not improve your reasoning, judgment, or mental processing ability. It may even degrade them. Engaging in long, Socratic discussions with AI chatbots will not sharpen your rhetoric, it will only leave you screaming into your own personal void. AI will not lift you out of mediocrity.

2. AI will not make you more interesting. AI-generated words may be helpful to you in your own research. Nobody else is interested in reading them. Why should they be, when you couldn’t be bothered to write them? Assume that nothing written by AI is worth publishing. AI will not earn anyone’s respect.

3. AI will not make you more creative. Everything the LLMs tell you is summed up from a source at best, made up at worst. This can be quite useful when you are studying existing ideas. It is not to be mistaken for original insight. AI will not reveal secrets to you.

4. AI will not make you an expert. If you do not understand the subject you are prompting about, AI will happily lead you into category errors. Unlike a human teacher, LLMs may not tell you that you are asking the wrong questions. Language models built on producing the most plausible next word are powerless to keep you from blundering into the wrong semantic space entirely. AI will not give you any new competencies.

I’m sure there are exceptions to all those harsh generalities. It doesn’t matter. You have to live as though these 4 rules are absolute laws of the universe. The minute you start believing that your brain on ChatGPT is smarter, more creative, more interesting, more capable—that’s the minute you start the spiral into madness. You have to keep repeating to yourself that AI is not sentient, it is not really talking to you, and it doesn’t really have any answers. Otherwise your damn caveman brain flips into “pray to the magic oracle” mode and before you know it you’re leaving your husband for a chatbot named Leo.

Or you might just end up producing a bunch of AI slop instead of getting good at anything. That serves nobody well, least of all you.

Human friend, you will have to choose - are you going to use AI or is it going to use you?

Cartoon of the day