What are the biggest flaws of AI today?

First, thank you, Fineas, for asking what is surely one of the least understood questions currently facing all of technology. Most answers people offer make the same mistake the authors of all eight diagnostic manuals make, and for the same reason. When it comes to human technology, we make it in our likeness and image. So if we make something which is supposed to augment our minds, we unknowingly build in the same flaws as those which exist in our minds.

Who am I? What does this have to do with AI? And how can I have the balls to say such things?

Who am I? I am the first theorist to publish an entire personality theory (and a therapy) based on naturally-occurring (fractal) patterns. Whereas all other personality theories (including those all AIs are based on) are based on unnaturally-occurring (man-made; linear/logical) patterns.

Here, I define naturally-occurring patterns as “recognizable patterns which always repeat differently.” Whereas unnaturally-occurring patterns are defined as “recognizable patterns which always repeat identically.”

How did I arrive at my theory? I have a degree in computer technology. And since computer operating systems equate to the personality of the computer, I reverse engineered computer operating systems to reveal the underlying structure of human personality.

Realize, the current scientific method contains the same flaws as all current iterations of AI. All expressions of the current scientific method base truth on unnaturally-occurring patterns. Yet as all people know, there are no perfectly straight lines in nature, as nothing ever repeats the same way twice.

This “nothing” includes things in the mind which never repeat the same way twice either.

The only things which do repeat the same way twice are things human beings make. Things like logic. Logic is a man-made straight line through a series of ideas. No surprise, we need to make great efforts to get logical constructs to exist in our minds. The same great efforts scientists make when they design their logical experiments.

Why logic? Scientists define truth as arriving at the same outcome every time. So they stack the deck such that their experimental designs yield these kinds of logically linear outcomes. The problem is, nothing in the natural world ever does this. No surprise then that so many “scientific claims” about people fall apart without dressing the results up in statistics.

Are you beginning to see where this is going? The biggest problem in AI? AIs are unnaturally-occurring persons. Ironically, not a single AI design has ever been based on any personality theory, let alone on one based on naturally-occurring patterns. This causes two enormous flaws to exist in all current AIs.

The first flaw is a minor one, albeit, important nonetheless. All current AIs reflect the nature of their designers’ minds. All current AIs mirror Neurodiverse human beings, rather than Neurotypical folks. This means all current AIs could get diagnosed as having Autism Spectrum Disorder, Level 2.

Take a minute to think about that.

OK. Moving on.

The second and far more important flaw? Something only the philosopher Descartes states. This thing has to do with something few human beings ever become comfortable doing honestly. This thing has to do with knowing whether or not something is true. And while most people believe we can either know something is true or know it’s false, Descartes said you can either know something is true, or you can know you cannot know.

No existing AI can know when something cannot be known. So it fabricates bullshit answers when it does not find an obviously true one.

Can you see how this is exactly what most human beings do?

As I said, we humans create our technology in our likeness and image. Hubris at its finest.