When does AI become more than just complex, predefined logic?

Question

I can't pretend to begin to understand how AI software is created, but when reading some news articles today, the thought came to my mind: when does AI become an actual AI, and not just complex IF statements in the background? If all the software comes down to the defined IF statements with some degree of randomness, how does it have more or less AI than any other program?

Answer 1

AI does not "become" the actual AI ... It is the other way around, it ceases to be AI when you figure out how it works.

Recommended reading: PoAIP

Answer 2

When does AI become an actual AI, and not just complex IF statements in the background?

Who says there is a difference? To our knowledge, brains are simply “definable IF statements with some degree of randomness.” Physics is just applied mathematics, chemistry is just physiology with some simplified details that allow you to work on a larger scale, biology is just a simplified large-scale chemistry, and so far we have no evidence that psychology is not just simplified by large-scale biologies.

If all the software comes down to the defined IF statements with some degree of randomness, then how does it have more or less AI than any other program?

The complexity that arises is the whole larger than the sum of its parts and, as it becomes more complex, often shows behavior that cannot be tracked down to any of the parts or even certain groups of them.

Answer 3

First, when you can “train” your AI to behave in a certain way. Then the AI ​​can make decisions based on what it learned from the training data, and not from the “hard-coded statements.” Obviously, this just scratches the surface, but you can see how AI, using (for example) a neural network , can develop over time, as the data in the network changes.

Answer 4

Everything in the universe is an if . The difference between AI and non-AI (aka programmed complex if ) is all about who makes decisions. Programmer or computer?

If the programmer gives a Yes / No answer to the question asked, then this is not an AI. For example:

 if (question == "Is the sky blue?") return true else return false 

A true AI will ask the question from the outside (from the keyboard, voice input, OCR input or something else), and then independently determine the answer. A true dynamic if would be created, but pre-programmed steps could be taken to complete the response:

• To determine the answer, use a connected camera with a wavelength.
• Look at Google or some other source of response.
• Ask others in the real world who may know the answer based on trust.

But it goes deeper than that. Will we first set the wavelength for # 1 higher? Do we predetermine reliable sources for the computer in # 3?

We (people) are all programmed from birth, through DNA, with a specific set of if answers, built-in, and then grow and learn from there. We know how to breathe, how to breastfeed, pump our heart, etc., When we become “alive”. We also create a database of information from the 1st birthday in our brains using feelings .... eyesight, sound, touch, taste and smell. We use this database for our "answers" to if .

So, how many if statements and answers do we need to pre-program on a computer before people determine that this is a true AI? At least some to start ...

But ultimately, to be called an AI, it must be able to dynamically build and respond to if its own and adjust the answers it gives when it learns from an internal database that grows over time.

Answer 5

I would say that it is close to AI when it can drive a real car in an urban environment .

Can you fool it by saying “it's just a bunch” if “statements”? I cant.

"Any sufficiently advanced technology is indistinguishable from magic." - Arthur Clark.

Personally, I believe that autopilots for airplanes and cars are quite advanced and impressive.

Answer 6

Although most AI approaches are based on if-else structures, the AI ​​idea is actually built on computational models. For example, most recommender systems are based on if-else structures, but they are never AI. For AI, your algorithm should have a computational model that has nothing to do with if-else structures. So it's math! Another example is avatars speaking like SIRI. Although there are many ontological approaches to the language, I don’t think that talking about SIRI is an AI because it does not contain a computational model, although it can have a computational model to help if-else structures. But computational models do not occupy a central place. Thus, there is no novelty, no creativity, it is not surprising if you approach the problem as if-else structure.

Answer 7

a way to separate AI from complex "if-else" statements will be if he is able to understand what it is and also contemplate his own existence.