Arguing with LLMs

Ever been in a debate with a colleague from a different role, and suddenly they drop a ChatGPT response on you to counter your point? 🤨

If so, you're not alone. It seems there's a new weapon in the arsenal of technical arguments.

We all love AI tools. They're amazing and help us with a variety of daily tasks with an efficiency that sometimes feels like magic.

But turning them into the ultimate authority in a professional argument? That’s where I think we’re missing the point.

Why do I find this so problematic?

An LLM is not documentation 📚
It hasn't read your internal design docs, it doesn't understand the project's constraints, and it can invent an entire API that doesn't exist.

An LLM is a confirmation bias machine 🧠
With the right prompt, you can make it say anything. "Write five arguments why X is better than Y" – and it will. That doesn't make it right or smart. Basically, it just wants to please you.

An LLM is no substitute for a real discussion 🎯
The whole point of a professional discussion is to weigh the pros and cons, to understand the trade-offs. When you pull out a generic bot answer, you're essentially saying: "I can't be bothered to think, let's just take the first answer from Google on steroids." You're just shooting yourself in the foot and missing an opportunity to learn.

I'm not against these tools, quite the opposite. But there's a huge difference between 'fake it till you make it' and 'let the AI fake it for me.' Using them to "win" an argument, without scrutiny and real references, is an abuse of the tool and cheapens the professional discourse.

So, have you encountered this phenomenon?
How do you think we should handle an argument that's based on a generic AI response?