10.15.2024.
There are many people warning about the dangers of AI, since there are many ways to misuse it. However, not many people talk about the spiritual dangers of overestimating AI by claiming that it possesses rationality.
Of course, as any religious person will know, humans have a soul, while machines do not. However, many non-religious people claim that while a program is electricity flowing through computer circuitry, thoughts are simply electric signals between the synapses of the brain. In order to respond to this, in the following paragraphs we are going briefly to discuss the purely materially observable, obvious differences between the “intelligence” of an AI, compared to the “intelligence” of a human.
The following arguments are based on Chapter 29 of “Only Man Bears His Image”. This book was written by Daniel O’Connor, an American philosophy professor and former engineer.
Differences between human memory and computer memory
The memory of a human resides in its soul, with the brain participating in the recall of the memories. This principle can be seen in cases where people suddenly remember things they forgot a long time ago. On a computer, the data is always stored on some hardware. If that hardware is destroyed, the data is lost forever.
AI cannot break the rules imposed upon it
This is a very important point. A human can break laws: whether they are laws created by other humans or moral laws. On the other hand, an AI can never violate its programming. For example, if it is written into an AI that it cannot scrap certain things from the Internet, or use certain words or give certain kinds of answers, it won’t do it. Because it can’t do it.
AI cannot commit mistakes, the results flow from its algorithms
People make mistakes all the time. A machine, however, always gives a result which is in accordance with its programming. A computer’s CPU always executes all code accurately. A “wrong” piece code is only wrong because it is wrong for humans. Programmers debug programs, since code can be “fixed” to give an output which is “correct” for humans. Simply put: AI is simply electricity running in a deterministic way through circuitry.
AI can calculate, but it doesn’t understand mathematics
O’Connor explains: “Consider this: whoever you are, I am sure you can easily add together any two integers. There is no limit to your ability to do this because you understand the nature of addition, and your capacity for applying that understanding is infinite. Since AI does not understand anything, including addition, it can only give the appearance of this understanding when individual ‘numbers’ (electrical flow which we interpret as numbers) are fed into its logic gates and combined. This means that it has no ability to even add numbers that cannot fit into the registers of the CPUs it operates on; whereas even a schoolchild easily could do this – for such a child understands addition.” (p. 582)
There is no algorithm which can generate actually random numbers
A person can easily name a random number. By contrast, a computer can generate random numbers in one of two ways. A “hardware random number generator” (HRNG) obtains “random” numbers by measuring some physical phenomenon. For example, it can measure random current fluctuations in circuitry known as shot noise. The other, more common, method is a “pseudo-random number generator” (PRNG), which runs an algorithm to get a number. This means that the number is not really random, as it is a result of a mathematical calculation.
Note: Daniel O’Connor does a great job explaining why AI isn’t rational and why aliens don’t exist. However, he also promotes false apparitions such as Medjugorje and Kindelmann, which we already debunked.