I've asked many people, "What is AI?" and the answers I get range from meaningless, to outright strange. Then they go into what AI will eventually do without really explaining what it is, or defining it.

So, what is it?

I know it's going to replace humans in some types of jobs. I get that it somehow involves a "machine" which will be "learning" things, whatever that means.

But when I say, "Oh, it's just advanced software," I get corrected and told, "it's not software, It's AI", that's where it gets a bit strange.

I'm also told, "it's not just algorithms, it's 'AI'". At that point, I realize they are using the term we are trying to define to define it.

In terms of the computers used, oh wait, I'm told they are "machines" not computers, but it looks like the same computer topology that's been used for decades, but I'm told, no, it's different than computer hardware.

I'm totally confused. What on earth is AI, and how does it differ from advanced software run on a traditional computer?

Finally, can anyone define AI without using the term AI?

Thanks, and sorry to be so confused.

It's Skynet.

Can't give you a spectacular definition, but can give you a great quote (though I don't have the attribution): "Artificial Intelligence is no match for natural stupidity."

In the early days all computer software was pretty much linear. You do this, then you do that. It was very predictable what was going to happen (at least in theory). As software has grown and evolved things have become more complicated. AI is not linear, and it involves the computer "learning" and making decisions based on what it has learned rather than only on what was programmed in.

Another term used lately is "machine learning". I dunno if this simply another way of saying the same thing of if "machine learning" is somehow distinct from "Artificial Intelligence".

It is simply multiple processes of computer programs that combined are able to do predictive, deductive and reasoned (logic) processing of data to provide "most likely/most correct" solutions to the "problem" the programs are designed to solve.

Complexly, it gets complicated.

ALEXA - What is

quote:

"Artificial Intelligence".

I would check out the YouTube channel 'code bullet' he does a lot of programming of "AI" to play computer games.

It is just algorithms and networks. You set a bunch of parameters and then set a goal or a couple goals and then the program runs hundreds of thousands of times figuring out how to accomplish the task. Its really just brute force of trial and error until it gets it right.

To me, AI and machine learning are just buzzwords people like to use to feel futuristic. Its just an program running a process over and over changing variables until it solves the problem.

Now, there are actual researchers on AI that I'm sure are legitimate. It as well as machine learning.

But my gut feeling in regards to the AI buzz-wording going around these days it just glorified rebranded skip logic / branch logic where all the scenarios/responses have to be pre-programmed.

I think most things currently in play are half-baked implementations from startups trying to get VC funding.

Or that is just the cynic in me.

20 or so years ago it was called Neural Networks but has evolved into AI and machine learning.

Machine learning is AI applied to a machine with a computer that causes the machine to perform some task based on measurements or observations by the machine using various sensors like visual, temperature, speed, and other measurable parameters.

AI can be purely neural network based but can also include heuristics in making decisions. Like “don’t hit the wall”.

Neural networks have to be “trained” by a large set of stimuli and correct responses. The stimuli would be the inputs from all the sensors. Responses would be the results of some simulation, model, or whatever (control laws, heuristics, etc.).

Huge amounts of training data is required depending on the complexity of the system. The AI or neural nets are computed from the training data. Then, when the system is deployed it can perform the correct tasks based on inputs from its sensors. In theory. AI or NN is only as good as its training data.

But, some AI or NN can “learn” beyond just the training data. So some machine or system might learn that if “X” happens do or don’t do “Y”. Who knows whether or not it will learn correctly - especially a complex system.

AI, ML, NN are buzzwords and technology coming back into fashion now that small, embedded computers have become so powerful and cheap. That is, these things can now be realized in a cost effective manner.

And thus, researchers are getting $$$ to develop, refine, and build systems with AI to solve some perceived problem. Like a self-licking ice cream cone.

Machines aren’t going to take over the world. The fact that they can learn doesn’t mean they will learn correctly.

Well it is still no replacement for the human mind. Thinking outside the box and creativity are just two examples of things AI cannot do. Human learning is not the same as a chess match.

My view, and it's probably not shared in the AI community, is that currently AI is best thought of as a kind of machine learning that takes the place of a degree of programming.

For example, my cell phone has facial recognition capability for unlocking. When I set it up, it took a bunch of views of my face to recognize me, and unlock. I did this without my reading glasses on, thus, each time I need to unlock my phone and I'm wearing glasses, I had to put in my PIN. Now, I'm starting to see that it recognizes me even when I have my glasses on. I suspect that there is some low level machine learning going on, such as, the camera looks and sees my face with glasses and rejects the login, but immediately, I put in the correct PIN and unlock the phone. If the phone now adds that "look" to it's database about my face, it will eventually "learn" that the face with glasses is an acceptable substitute for my originally programmed face, and thus it will "learn" to unlock my phone without me consciously doing additional programming to add a face with glasses.

This type of AI or machine learning is essentially the machine being programmed to add certain things to a database such that it makes it's predictions from a larger data set that is not purposefully programmed by a human, rather a human programs the machine to add this information to its decision matrix.

True AI would be the machine understand an if-then equation without being instructed to do so. We are a long way from that point, even further from a machine understanding an if-then operation in context of a larger set of unprogrammed factors, and incomprehensibly far from ZSMICHAEL'S point about creativity.

A natural blonde, dyeing her hair brunette.

It's also what they are calling "education" in today's college system

quote:

Originally posted by RichardC:
A natural blonde, dyeing her hair brunette.

Winner Winner ...

Good example of AI is a Nest Thermostat, it "learns" your habits. It is merely software that can recognize trends and that "adapts" to those trends. The problem is that we use terms like "learning" and "intelligence" and apply them to software. Like network switches that we call "dumb".

Now, quantum computing and AI will take you down a rabbit hole that even one of my best friends who is a physicist gets lost.

I want my flying car.

This is an example of the technology as an actual deliverable:

https://ir.rockwellautomation....uction-/default.aspx

As far as I’m concerned, it’s not AI until it can pass the Turing Test. Repeatedly (to account for examiner or other structural bias in the administration).

-Rob

My youngest son has been at this stuff for several years, recently selling his company to a large contractor. His R&D position with them is a big step, one I neither understand nor have the attention span to try

quote:

Originally posted by ZSMICHAEL:
Well it is still no replacement for the human mind. Thinking outside the box and creativity are just two examples of things AI cannot do. Human learning is not the same as a chess match.

It doesn’t sound like you are referring to a true AI. A true AI is more than learning or adapting. It has to be self aware. Siri and Alexa are not even close. The ability to reason is also not there in current “AI” systems.

Robins lay eggs. Check.
Birds lay eggs. Check.
Therefore, robins are birds. Nope. Most systems just aren’t at this point yet.

quote:

Originally posted by sigmonkey:
It is simply multiple processes of computer programs that combined are able to do predictive, deductive and reasoned (logic) processing of data to provide "most likely/most correct" solutions to the "problem" the programs are designed to solve.

Complexly, it gets complicated.

I'm still lost on how AI would differ from traditional software. Let's take, as an example, the alarm system on my house. It has several doors to monitor concurrently, so that's multiple inputs. I sometimes enter through the front door, and sometimes I enter through my back door or side door.

If I was to ask the computer running the alarm system: At 4 PM next Saturday, which door am I most likely to enter through? The computer simply looks at the door open/close log from the past, and sorts out the Saturday afternoon events, and realizes that on Saturday I use the rear door more often than other doors. So it guesses rear door. But wait, I also have the weather logged too, and it finds from that data that when it rains, I use the side door more often (so I don't muck up the carpet near the front or rear door). So, the computer looks (on the web) at the weather prediction for Saturday and realizes it's gonna rain, so it changes the guess to the side door.

That seems like an AI activity, but honestly, I could have written this code on the IBM-XT clone I had 25 years ago.

I still fail to see what's really new about AI compared to what we were doing 25 years ago. It's just that now we have faster computers and more memory, and maybe more sources of data to look for correlation with. I just don't see how AI is somehow different. --please help me out here.

Also, Is the hardware for AI different, or is it the same stuff we used 25 years ago, only a faster version with more memory and more data inputs?