A lot of people you meet have decided AI is garbage — that it makes things up, steals from artists and writers, and that plain old Google works fine. If you actually use AI and want to say something, the worst move is to argue like a salesperson. Here’s how to talk about it without sounding like an ad.
Start by agreeing with the parts that are true. Because they are.
- It does make stuff up. AI confidently says wrong things. Anyone who trusts it for facts without checking is going to get burned eventually. That complaint is fair.
- The training-data and copyright stuff is a real mess. These systems learned from huge piles of other people’s work, and whether that’s okay is genuinely unsettled — legally and ethically. Pretending it’s all fine is exactly the hype that gave AI its bad name.
Saying that out loud makes you sound like a realist instead of a fanboy, and people actually listen to realists.
Then reframe it honestly: it’s a different tool, not a better Google.
- Search is the right tool when there’s a specific fact or trustworthy source to find. Keep using it for that.
- AI is better at the things search can’t do: pulling together a bunch of messy information, rewriting or summarizing text, drafting a rough first version of something, or helping you think through a problem when there’s no single page with the answer.
The pitch only works because it’s narrow. You’re not claiming AI wins at everything — just that it’s good at a few specific jobs you can double-check yourself.
And say the quiet part out loud: it’s totally fine not to use it. Plenty of people don’t need AI and lose nothing. It’s not a moral obligation, and someone happy with search and their own brain isn’t missing some essential thing. People dig in harder when they feel pushed — and get curious when you clearly aren’t pushing.
A one-liner you can actually use:
“Yeah, it makes stuff up and the training-data ethics are a real mess, so I don’t trust it for facts. I use it for drafting and thinking through things where I can check the output myself — and I still Google when I need a real source.”
That’s something a skeptic can hear, because it isn’t trying to win. And if they’re still not interested? That’s a completely valid choice. You don’t have to convince anyone.
Sources:
- OpenAI, “Why Language Models Hallucinate” — the company that makes ChatGPT explaining why these systems confidently produce false statements, and that the problem isn’t fully solved.
- Norton Rose Fulbright, “An update on AI copyright cases in 2026” — a law-firm rundown of the major AI training-data lawsuits, showing the legal questions are still unsettled with no nationwide answer yet.