AI is not the problem. We are

I've noticed it even on this forum. More and more people are blindly trusting the output of an LLM for theological arguments and spiritual guidance. In one case recently someone "refuted" an article by asking Chat GPT to do so and then just copied the section headings and clipped out the bulk of the response, ironically leaving only a set of assertions with nothing to back then up, but the fact that it came from AI was seemingly enough.

AI is a tool, but the problem with it is that it is able to output nonsense at a faster pace than human beings can counter. Existential crisis, I don't know. But certainly an epistemic one as our ability to verify trustworthy sources is eroded as we are flooded with AI generated garbage.

This reminds me of when pocket calculators became readily available to students.

Decades ago, I wrote a program for work, and when another tech saw it, he said, "That's a good dodge. They'll believe it (the results) because it's a printout." I think he was right. That's back when AI was basic neural nets, not really used for anything. AI is another version of "It came from a computer, therefore it must be right." Meanwhile, Garbage In, Garbage Out is still as much as a thing now as it was then, and just because it came from a computer doesn't mean a thing other than it came from a computer.

Now we have AI, which I don't know how it stacks up compared to, say, a cockroach's brain, but it's the latest and greatest and people are turning to it for all sorts of things, never asking if garbage was used to train them. It's a threat only in how much we let it be. Given AI any autonomous control over, say, weapons, is a very bad idea, not because of a Skynet scenario but because the things are not really predictable. Some SF story I read decades ago had sentient warbots that went from targeting humans in a specific uniform to targeting all humans. Now, if they were sentient, they would have known a little of why they were supposed to go after specific humans, but the not very intelligent AI we have might make such a leap.

AI generated text has been a problem because of human laziness. SF/F publishes have already been hit with AI generated stories to the point where they state they won't accept them. Some call it an issue with cutting and pasting from existing works, which brings up the issue of plagiarism. AI school papers come from the same urge as using notes of literature rather than tackle required reading, but are a problem. Interestingly, the response has been a return to cursive writing and handwritten assignments, preferably done in class.. Kind of hard to have AI write your paper if the professor is looking over your shoulder.

And now, a word about calculators.: They have had a negative impact. The most obvious is running out answers beyond a usable number of digits. Another is the lack of ability to do a mental estimate as a check. Not guessing, but running a close enough calculation in your head to check results. Fractions are taught but the use is unfortunately lost. Yes, the "good numbers" like pi and e and the square root of three are irrational, but expressing some calculations as fractions lets them be done more easily mentally. Such as:

My speedometer says I'm going 55 mph but am I? Have a chrono watch but it's a bad idea to pull out a calculator while driving to figure how long it would take to travel 1 mile. But 60/55 reduces to 12/11, which is 1 1/11 minute per mile, which is 1 minute, 60/11 seconds, which is 1 minute 5 5/11 seconds, so if it's between 1 minute 5 seconds and 1 minute 6 seconds between mile posts, it should be good.

Did I mention I'm bad at math? Had the grades to prove it. So if I can do this mentally, it's not that hard. But if someone is slow with mental math, like me, and tries digits, it's a huge slowdown. It can be done, but it's much slower, at least for me.

Tuur said:

Decades ago, I wrote a program for work, and when another tech saw it, he said, "That's a good dodge. They'll believe it (the results) because it's a printout." I think he was right. That's back when AI was basic neural nets, not really used for anything. AI is another version of "It came from a computer, therefore it must be right." Meanwhile, Garbage In, Garbage Out is still as much as a thing now as it was then, and just because it came from a computer doesn't mean a thing other than it came from a computer.

Click to expand...

With AI there's more than just garbage in, garbage out. Even well crafted queries are likely to produce garbage because AI is engineered to give "pleasing" results rather than accurate ones. So if there is nothing that would please the person making the query it will produce false results, and make them appear as convincing as possible. So it's basically a recipe for perpetuating confirmation bias at a rapid pace.

AI hugely helped me land a decent job opportunity I never had for many years. I was getting suicidal already. It helped me compose the replies and prepared me for tests and interviews.

Not that I didn't asked other people for help first. I did but none was able to help me with what I really needed.

Michie said:

Alongside those who think AI will save the world and revolutionize everything are a growing number who think it will destroy it, or at least come close. In a recent episode of Ross Douthat’s “Interesting Times” podcast, former Open AI researcher Daniel Kokotajlo warned that Artificial Intelligence will become an existential threat to humanity within two years.

https://www.breakpoint.org/ai-is-not-the-problem-we-are/

Click to expand...

I've been testing chatgpt extensively. I agree with the openai researcher.

However I don't think the real threat is to humanity but to capitalism and those who will sacrifice their lives to defend it.