# AI, Rocks, Clever Hans, and Analytics By:: [[Brian Heath]] 2023-05-23 A few years ago, everyone would talk about machine learning as the pinnacle of analytics. Today people are moving on to talk about Artificial Intelligence as the pinnacle. These technologies and tools enable types of analyses, but they are not analytics. Machine Learning is an algorithmic technique closer to a statistical regression model than any learning. Artificial Intelligence is not intelligent. It is merely a clever magic trick to reproduce what has already been done in new ways. Neither technique thinks or understands. It only processes what it was programmed to do. Skeptics may argue that humans are no different. We take inputs and produce outputs just as these algorithms do. If you believe this means Machine Learning and Artificial Intelligence are on par with humans, you must also believe that a rock is on par with humans as well. After all, a rock takes inputs such as someone hitting it with a hammer, and produces sand as output. Who's to say that a rock isn't thinking and responding? Who's to say that the rock is content doing nothing all day? In Computer Science, the Turing Test is used to see if a human can distinguish whether responses are coming from a machine or another human. If they can't tell the difference, the machine passes the test of being sufficiently "intelligent." However, this is a low bar for intelligence, and humans are horrible at objective evaluation. People have believed in many crazy situations where odd things seemed intelligent. For example, in the early 1900s, a horse named Hans gained worldwide fame as a horse that could do math. People would ask Hans math questions such as adding two and three. Hans would stomp his hoof five times in reply. After much study, it turns out that Hans was reading the body language of his trainer to know when to stop stomping his hoof. The same parlor tricks play themselves out today but under the name of Artificial Intelligence and Machine Learning. They are useful tools but they do not think or even come close to it. Analytics is about thinking and understanding. Unlike Hans, the goal of analytics is not to stop doing analyses simply because the manager gets the answer they want. Machine Learning, Artificial Intelligence, and Hans all stop on an answer when asked. Analysts have the option, freedom, and obligation to continue at their own risk. #### Related Items [[Analytics]] [[Artificial Intelligence]] [[Machine Learning]] [[Intelligence]] [[Thinking]] [[Magic]] [[Beliefs]]