In my post The End of Artificial Intelligence I raised the question “What is real intelligence?” I went on to refine this question to “What criteria constitute intelligence on the same level as we humans?” This is not an easy problem to address, primarily because not everyone means the same thing when referring to the quality of intelligence and it can be used to mean different things in different contexts. For example, if a person makes a lot of mistakes we might declare that they are not very intelligent. If a person is very good at their job, especially when that job requires learning something complex and some degree of mental difficulty, such as solving problems or managing a financial portfolio, we refer to that person as smart, astute, or intelligent. Yet, if I were to create a machine that is capable of doing that same task, managing a financial portfolio for example, most of us would hesitate to say that this machine was intelligent. We might say it a good system, perhaps even advanced. We might say things like “this machine is very capable.” But to confer intelligence on a machine is not a leap most of us are comfortable making. So exactly what is it we are looking for before we can confidently pronounce a machine to be intelligent? I will attempt to lay out a few basic criteria. Whether these criteria are all necessary for intelligence I leave as an open question along with whether these criteria are sufficient for intelligence. For the moment, my goal is only to come up with some minimal and rather basic notions for consideration. In other words, where does intelligence begin? At what point are we willing to say that a machine possesses intelligence?

In my next three posts I will present the things that I feel are probably the minimum characteristics of any intelligent entity. This will form a basis for discussing how to apply those criteria along with caveats, concerns, and other possible factors to consider.