Ive noticed a lot of good people take awful political positions this way.
Usually they trust the wrong person - e.g. by falling victim to the just world fallacy ("X is a big deal in our world and X wouldn't be where they are if they werent a decent person. X must have a point.")
Yeah, with drugs being legal now I've been wondering what the next boogeyman we're going to declare war on would be.
To do that level of research, one needs strong independent thinking skills and deep technical expertise. We have plenty of evidence on this very site how hard it is to influence technical people and change their opinion.
However, you seem to be attributing super-human capabilities to a whole group of people for some reason
Yes, the current version of AI is not capable of large scale harm by itself yet, but the plausible trajectory is worth warning about. Gordon Moore did make fairly accurate predictions after all.
You are right. And no one is arguing the contrary
We can all see the risks
The thing is, what do we do about it
Regulating AI will only serve to help the big players and keep others out
I do agree that the likelihood of regulatory capture in this arena is very high.
In most projects, many unmatured software is being used just because works for some specific task, nowadays.
With that in mind, it is easy to hyphotetize that many projects are already using LLMs internally. Not always for good deeds, easily, one of the most and best use cases, is to use LLMs to command & control distributed malware.
So there you go, it maybe possible that some beta level of intelligent malware is already roaming the Internet right now. We'll know for sure in some years from now (the usual time for advanced malware to be discovered is somewhat 2-3 years after it has been took it to production).
They might not be getting paid, but that doesn’t mean they are not being influenced
AI at this point is pretty much completely open, all the papers, math and science behind it are public
Soon, people will have advanced AI running locally on their phones and watches
So unless they scrub the Internet, start censoring this stuff, and pretty much ban computers, there is absolutely no way to stop AI nor any potentially bad actors from using it
The biggest issues that we should be addressing regarding AI are the potential jobs losses and increased inequality at local and global scale
But of course, the people who usually make these decisions are the ones that benefit the most from inequality, so