Confidence changes how people evaluate information…[P]eople tend to use confidence as a shortcut for assessing credibility, especially when accuracy is hard to judge. The effect persists even when people know a system or person can be wrong.” — Nick Dothée in The Atlantic
AI remains a locomotive traveling at high speeds through a dark tunnel without a conductor. We don’t know yet how things will shake out on the other side. The train may not even stay on the tracks at this speed and no guardrails.
Wherever happens, the important thing to remember right now is to tread with caution. If you use AI tools, like Claude, ChatGPT, Gemini or other “advisors,” there’s a right way and really bad way in how you use them for information retrieval. Not talking about coding or any number of specialized tasks.
Dothee says what keeps him coming back to LLMs is not their accuracy. We all know they hallucinate and give wrong, nuance-free, and incomplete answers. What keeps him coming back is ChatGPT’s confidence. The platform replies with such authority and swagger and even keeled tone, sounding like it knows what it’s talking about.
That’s a problem because it starts to corrode our own confidence.
The same thing is true with human beings. We all know someone who thinks they’re always right and never shows doubt or qualifies what they say. Falling for someone’s faux confidence, especially at work, can boomerang back on you if you’re not discerning. Some people take “fake-it-until-you-make it” too far. Or they never make it.
Moral of the story: Don’t turn off your B.S. detectors when you go on AI platforms. That little voice inside your head should be constantly questioning how the AI bot answered your prompt, even with a really strong, specific one. Make AI bots provide their sources. You are the fact checker in control—not the AI.