AI and customer profiling.

The AI can predict things and act before a person can even open their mouth. This is possible in cases involving individual people. And specific locations, such as certain restaurants. The system must only have a profile of the person, and then it can predict the person's orders. The system can predict the portion. What a person orders. If. It has knowledge of the person’s favorite food. In a limited space and a limited number of actors and choices, the AI can make predictions very easily. In the case that the system knows. If a person hates boiled cabbage, it can make excellent precision. 

If all portions except one. Involves baked cabbage. The system can make a profile. Simply by asking a person about food. The person will not eat under any circumstances. Or the system. It can follow what food the person throws away. But. There is always a possibility that a person will order the portion that involves baked cabbage anyway. If other things on the saucer taste good. 

These kinds of things make the prediction. For a single person. Almost impossible. We can predict. Somehow, how large groups of humans behave. 

We have scientific models. Of. Behavior in certain situations. We can predict how atoms behave in large-scale models. But we still cannot calculate the point of the single atom in the room. 

We can predict how animals behave. We know when birds make their descendants. But the thing that separates us from animals and atoms is that we can suddenly change our minds. This makes this kind of prediction hard. All people. They will not react the same way. In the same things. Things like our experiences and many other things determine how we react. If we sit in a restaurant. We might not buy food for a garbage can. So, if we throw the food. That we paid. Into the garbage, that means we don’t like it. 

For making profiles, the system can follow. What bites the customer puts on the edge of the saucer. This kind of prediction is possible. But. If the AI wants to predict how a person behaves in normal situations, that can be difficult. In cases like stock marketing, the system must only know whether the stock value rises or not, and then it can make a prediction. About how the investor behaves in that area. The idea is to bring profits, and that means if the stock value decreases, the system must not have a large-scale imagination. 

Then, when the stock value decreases, that means the actor sells those stocks. Or they should sell them. If they want financial benefits. This thing requires that the system can handle two variables. 

But then we come to the cases that we call human. Humans act differently from what nobody predicts. We can have the same morning routine every single day for years. We can come to the same bar. Or cafeteria. Every morning for years. The staff might know us, and maybe they know what we will order. But then one day, we will not come to that cafeteria.  We find another place. We might move to another city without warning. And that makes. It's hard to predict. On how a person behaves. In a large-scale environment. There are many variables that the system must notice to predict a person’s behavior. It’s almost impossible in a large-scale environment. 

https://bigthink.com/science-tech/proactive-ai/