Artificial Intelligence

[Arran]

Not sure it's do-able?

Seductive as the thought may be, I'm not sure a truly humanlike thought 'engine' can be produced artificially. We'd need to understand - in excruciating detail - how the entire brain works, what consciousness is, what emotion is, what judgement is, what feelings are, etc., etc.

And so I wonder if the human brain's just too complex to be understood. Maybe we can understand a simpler brain (like a fish, or a mosquito) in enough detail to build a functional equivalent? I'm just not sure we have enough brainpower ourselves to fully comprehend what makes us 'us' - if you know what I mean.

Emerson Pugh said this more elegantly and efficiently: "If the human mind was simple enough to understand, we'd be too simple to understand it."

 

[Jaro65]

An aspect of this question that I find interesting is whether we will be able to accomplish AI using the traditional binary logic, or if some other approach (fuzzy logic, or something not invented yet) will be required. I attended a lecture by Dr. Zadeh, the inventor of fuzzy logic, some years ago and even he didn't know if current technology can be used to implement algorithms that would successfully imitate brain activity. Of course, we are now assuming that human reasoning and associated aspects of continuous learning, emotion, morality, etc. can actually be coded in some type of algorithm (regardless of its complexity).

But then again, our grandfather's never even dreamt of a pocket calculator....

 

[Huntn]

Thread revival!

Why start a new thread? I'm involved in an intriguing discussion in this forum about a recent movie concerning AI, Ex Machina. Please use spoiler tags or warn of spoilers if you mention the plot of that movie, thanks! Anyway...

Anyone working in the field? What is the current state of Artifical Intelligence? While there might be a computer capable of passing The Turing Test, this is much less ambitious than creating an android, as in the movie that could pass for human. According to this BBC article from 2014:

Computer AI passes Turing test in 'world first' (2014)

Sounds like this claim was disputed, but the more I learn about the "test" I realize how limited it is, what a low bar, text chat versus face to face interaction. We'd need an expanded Turing Test for androids.

The question for androids emulating human beings, would we ever want something that could pass for a real human or would most of us want an idealized human, lol? It's one thing to display intelligence, logic, reasoning, personality, communication, but besides The Three Laws of Robotics, you'd need a morality routine with hard limits, (such as The 3 Laws), and you'd want to allow positive emotions but dampen negative emotions. After all, we want happy androids don't we?

This seems to be the cutting edge, but the challenge seems huge and androids still have a beyond the horizon, long way to go.

Human or Machine? Life-Like Android Robots from Japan Show Glimpses of the Future (2014)

​

Looks good, but can she speak and move looking realistically human?

 

[Huntn]

I'm curious if there have been any programs marketed that allow the creation of rudimentary AI programs without having to know and write basic code? Maybe something like that Apple scripting program whose name eludes me.

 

[Huntn]

Never mind... Please delete.