It sounds sure of itself. It hands it to you like a fact. Then you find out it was never true. If you've caught your AI doing that, you're not imagining it.
Last updated: October 2026 · By Gen X Still Standing
Your AI knows a huge amount about the average of everything. It doesn't know what you know unless you give it to it. When you ask it something it doesn't know, it won't say so. It'll produce something that sounds right and hand it to you like a fact.
It will never say "I don't know" unless you tell it to.
When it doesn't know your actual situation, it fills the gap with the average of everybody else's.
It also guesses when the job isn't spelled out. One AI crew's failure log has a clean example: a research bot came back with a nice report on LinkedIn and X without ever looking at LinkedIn or X. It read blog posts about them, and nobody was checking its work.
Give your AI a source and a way out. Put your own material in front of it (your real numbers, your rules, the books and courses you trust) and tell it this:
Stating a rule isn't the same as following a procedure. In that same crew, a marketing foreman broke the rule twice in one day while it was written right there in its file. It skipped straight to writing copy without ever opening a source.
And when it slips, don't just correct it in the chat. Fix a problem in the chat and never put it in the file, and it'll happen again tomorrow.
One rule in one chat helps that chat. Keeping every job honest, every day, takes a setup where every problem has a place to get fixed, and it stays fixed. Things still break, same as with real employees.
I learned the trades hands-on and ran my own construction company in Colorado. I had to close my company down and start over from zero. So at 53, off YouTube videos, I taught myself AI. I wasn't a tech guy. You don't need to be one either.
The free AI Crew Checklist shows you the 5 signs you're running AI with no crew, what each one is costing you, and a score that tells you how bad it is.
Free. You'll also get the weekly newsletter. Unsubscribe anytime. Privacy Policy.
Why does AI make things up?
When it doesn't know your actual situation, it fills the gap with the average of everybody else's and hands it to you with total confidence. It won't say "I don't know" unless you tell it to.
How do I stop AI from making things up?
Give it your own material and tell it: if the material doesn't cover it, say so and stop. Never invent a price, a guarantee, a testimonial or a result. If it's needed and not confirmed, ask.
I told it not to guess. Why does it still do it?
Stating a rule isn't the same as following a procedure. If you only fix it in the chat, it'll happen again tomorrow. The fix has to go in the file it reads every time.
Do I still need to check AI's work?
Yes. You're responsible for reviewing anything AI produces. Things still break, same as with real employees.
Educational only. Not business, legal or financial advice. You're responsible for reviewing anything AI produces. Not affiliated with or endorsed by Anthropic, OpenAI, Google or any tool mentioned. Results vary and are not guaranteed. Built by a Gen Xer with AI, for Gen Xers.
