Member-only story
Grok Is An Early Sign of AI’s Ability to Mislead
Musk’s chatbot may just be a clumsy version of the future
A few weeks ago, I was in a big box store in search of a very specific type of hose. As happens to me so often at these types of stores, I stood in the aisle, gazing forlornly at the vast array of products, suddenly less sure of what I wanted than when I left my house.
In this instance, I was joined in the gardening department by an outgoing woman who, it turned out, was searching for a similar sort of thing. We got to talking, and hit it off — it turned out that we had the same job (we are both teachers), and had similarly incomplete understandings of what we were trying to install in our gardens. Because this is 2025, we both whipped out our phones and started searching for answers, comparing what we found.
In the end, I didn’t find what I needed, but it was a fascinating experience to see the difference in the way we looked for information; the woman relied entirely on ChatGPT, while I used old-school Google. She told me that she “loved” ChatGPT and used it “for everything.” She proceeded to ask the chatbot all sorts of questions about the best type of irrigation hoses to buy and where to find them in the store.
