More A than I

I am not afraid of artificial intelligence. It is artificial. It is not real even though it gives the appearance of being real. As long as we don't fall for the facade and are not fooled by the superficial we should be ok. In fact, the best thing fro us to do is to quantify what a computer CAN'T do and we will be able to spot the usurper.

Fact -- if I populate the internet with writing that is intentionally incorrect, and use a specific word repeatedly, a scraping, generative AI system will, not knowing or caring that what I wrote is factually (or even logically) wrong. It won't even really care if what I post is syntactically incorrect. The fact that my writing exists will impact (even if only slightly) the statistics which  a system uses to determine what a potential "next word" might be. If I do it enough, or pick an esoteric enough word, I can make a more substantial impact on the statistics and someone, somewhere, relying on a generative engine, will end up with something ridiculous. And the computer won't understand what the problem is.

There are things that make us human. Body language, tone of voice, context, personal vocabulary, personal preference and affectations. These are things that can not be replicated by computers because, while they can build a database of foundational information, they can't choose to ignore it for no apparent reason.

Computers can't be illogical or alogical (arational vs. irrational). They don't have gut feelings and can't be swayed by emotion. This may seem like a good thing, but it isn't. There is no appeal that might sway a computer. They can't decide based on intuition or be intentionally wrong. They cannot act on a whim. If one understands their programming, one can anticipate their behavior. Even their "randomness" is not random. They cannot capriciously choose; they can only create the appearance of randomness.

Could artificial intelligence actually exist? Yes, but under our rules of understanding, it wouldn't be intelligence. Knowledge is easy. But intelligence is about the human insight into when to be smart and when not to be. All of the AI versions I have seen in movies are something fantastical because they transcend programming and require a leap that endows on technology things like uncertainty, emotional investment or other intangibles which are what make humans human.

So a computer can imitate a human. A computer can do all the things that one would think of as being endemic only to humanity but it is all just cosplay. In the same way that we can copy the technique of Jackson Pollock, but we can't innovate his approach because we are just copying what is already there. A computer can pretend or copy but it can't create. A computer might even be able to consider the variables which would help it come to a solution to a problem but it cannot understand why it might not want to weigh a specific variable the same as others in an exceptional case. It cannot intuit and go against the statistical evidence/precedent.

So I'm not afraid of AI. I'm afraid of the non-intelligence that people will (and do) mistake for intelligence. I'm afraid that the level of actual intelligence is so low out there that people are willing to accept computing engines following programming as intelligence.