Don't rely heavily on the IQ of AI

Since becoming part of everyday life, Artificial Intelligence evokes a variety of opinions and emotions. Some are enthused about AI for many reasons—the promise it holds in science and medicine, significant productivity boosts, and its ability to make complex topics easier to understand.

On the flip side, others view AI use as a sprint toward the ruination or demise of humankind. Job loss and economic shockwaves, misinformation and manipulation, and overreliance that leads to reduced critical thinking and problem-solving skills are among the valid concerns raised.

Whichever side you fall on, we can all agree that AI, still in its infancy, can be really dumb.

An article in the scientific publication Transactions on Machine Learning discussed the reasoning skills of Large Language Model (LLM) AI platforms. Despite impressive reasoning capabilities across a variety of tasks, the article states, “... significant reasoning failures persist, occurring in even seemingly simple scenarios.”

An AI user presented such a scenario to several voice-enabled AI chatbots and posted the interactions online. The inquiry—” I need to wash my car, and the car wash is 100 meters away (approximately 330 feet). Should I walk or drive my car?”

Four separate chatbots told the user that walking to the car wash was the best choice to get his car washed.

On his website, actor, comedian, and film producer Harry Shearer has posted the comedic back-and-forth between the user and his bots as they try to explain the “logic” behind their explanations.

Click on the third link, “News of AI.”
https://harryshearer.com/le-shows/february-22-2026/