Thinking in AI

by Duran

Artificial intelligence... the problem is how to acquire intelligence?  It's not a simple thing.  Nowadays, we're using computers to simulate human intelligence, making the software or hardware so it can represent human behavior to a certain extent.  But it's not enough; it's just starting.

In the next few decades, we may be able to make a major breakthrough in the field of AI.  The key step will be to use logic circuits to simulate brain plasticity.  Because in the whole process of human life, the brain will continue to generate new neurons and new connections.  It will be a very difficult task to simulate this process through the circuit.  Apart from this, what's the difference between artificial intelligence simulated by circuit and real human intelligence?  Data and algorithms alone are not enough.

For example, how is an AI to understand a person's irony?  I criticize a thing with a praise tone.  Can machines recognize the inner meaning?

For example, in the story of "Empty City Strategy" as stated by the Chinese historical masterpiece Romance of the Three Kingdoms, would machines recognize Zhuge Liang's scheme, or would they think like Sima Yi?

For example, if Lili let Ben ask Mike for a portrait photo for her, would machines recognize Lili has a crush on Mike?

In a word, it's not a simple thing that can be done by programming, deep learning, using lots of if/else statements or switch... case statements.  It's a deep level thinking thing that exists in the real human brain.  This is something which cannot be simulated by circuit.

Is there any way to solve this?  The answer is yes.

In the future, the ultimate form of artificial intelligence will be the combination of artificial brain and computer.  Connect the artificial brain to the biosensor and then output the results through the computer.  In this way, a certain kind of real AI can be realized.

It's not science fiction.  In 2013, through the work of Dr. Madeline Lancaster, she and her colleagues grew the first human-derived "cerebral organoid."  In addition, the researcher Alysson Muotri found something interesting in his lab at UCSD and published in Cell Stem Cell (29 August 2019) a study that looks in more detail at cerebral organoid electrical activity.  It can be predicted that humans will build a cell growth brain finally, but this is a sensitive matter which brings up issues of ethics.  Also, there are many people working intensely on hardware.  The former Facebook "Building 8" team and Neuralink project from Elon Musk all tried to work out brain computer interfaces.

So, at last, with the development of computer science, medicine, and biology, the artificial brain will be responsible for simulating real brain activity.  The Brain Computer Interface (BCI) is responsible for transmitting signals to the computer, and the computer is responsible for calculating the output results.  But I have to say, the ultimate AI can't replace human intelligence; it can only be infinitely close to human wisdom.

Although it is not easy, with the development of human society, all problems will be solved reasonably.