Chatbots are designed to please. You may have noticed that the minute you ask a question or share your idea, any given chatbot might come back with something like, “That’s a great idea!” If you ask me, it’s wildly unnerving. Not every idea is great and I’ll admit most of mine are very far-fetched, which is why I want a chatbot that can call me out, not stroke my ego.
I have tested chatbots’ varying excitement for my ideas, including a recent Tater Tot Cheesecake. But, if you lean into a chatbot that is willing to entertain every idea without any pushback, you’re going to get mediocre results.
That’s why after months of testing chatbots for everything from research to productivity workflows, I've found something surprising: AI gives significantly better answers when you challenge it. The best results don't come from the first prompt; they come from the first argument.
I’m not talking about being rude or aggressive, although I hear that works, too. But that’s just not in my nature; I’m talking about challenging it critically. Think of it less like arguing and more like pushing a smart coworker to defend their thinking.
