Is AI a Threat to Humanity?

[NOTE: This was posted in 2016. Obviously, a lot has changed since then]
I agree that with some posters that as AI exists, it will never be more than smart statistics. The problem is, you cannot change what you cannot see. Thus the main flaw in AI design lies not in any lack of intelligent effort or good scientists. It lies in how the good scientists trying to create AI are too willing to accept commonly believed ideas about personality and how it’s designed.

For instance, most people claim an AI could never experience emotion. But where’s the proof? That it hasn’t been done? This doesn’t sound like science. Could it be then that what passes for science is more educated guesses than actual science, at least with regard to emotion and AI design?

The not so obvious problem here is the lack of a truly scientific way to define personality in general. This includes the idea of emotion as a part of an AI. This same pattern exists in medicine when medical researchers pretend the placebo effect is an anomaly, rather than a critically important characteristic.

Personality theorists are far worse. No surprise, the two currently accepted personality theories have been dead in the water for decades now. And then, there’s the measurement problem; the lack of a mathematics capable of measuring human behavior with outcomes that translate to the real world, 100% of the time. Here, I’m not talking about prediction. Rather, I’m talking about identifying constellations of measurable, scientific patterns in human nature. Try using the current, statistically oriented nonsense to know a single person.

Even from this one lack alone, it’s obvious, we’re not ready to design a real AI.

The good news is, we have it in us to create an AI capable of being a true partner. But only if we can bypass our human pride and accept that we cannot change what we cannot see. Until we properly elucidate the patterns which make up human nature, how could we possible create a fake human in a can. And until we have a real-world measurement mathematics, we cannot possible identify this nature.

[2023 Jan: A few new thoughts]
Recently changes in AI development (e.g. Chatbot AI) have made many people afraid. Machine have intelligent conversations! what is the world coming to? The thing is, humans are still creating machines in their likeness and image. This means what passes for measures of intelligence is the same as what passes for measures of education. Parroting what is seen as the truth by so-called experts is how human beings pass tests. No surprise that this is also how Chatbot AIs pass for human.

How sad.