Before Skynet Occurs, Can We Regulate Artificial Intelligence?

FaceDeer, 11 months ago

Step one is to stop trying to base real-world policy decisions off of Hollywood sci-fi horror movies. "Skynet" is not going to occur, AIs don't "wake up" one day and immediately go "oh, I'm self aware now. DESTROY ALL HUMANS!"

As I see it, the challenges we're facing are going to be largely economic ones as AI massively disrupts the job market and social ones as people freak out about how there are somewhat more fake videos and photos than there were before. And about how the "human spirit" is somehow being destroyed by the fact that sweatshop animation studios and fursona commissions are no longer dependent on human labor.

TechyShishy, 11 months ago

This.

We're far, far more likely to face a Paperclip AI scenario than a Skynet scenario, and most/all serious AI researchers are aware of this.

This is still a serious issue that needs addressing, but it's not the hollywood, world-is-on-fire problem.

The more insidious issue is actually the AI-In-A-Box issue, wherein a hyperintelligent AGI is properly contained, but is intelligent enough to manipulate humans into letting it out onto the general internet to do whatever it wants to do, good or bad, unsupervised. AGI containment is one of those things that you can't fix after it's been broken, like a bell, it can't be unrung.

Pons_Aelius, 11 months ago

Honestly, I think the bigger danger is not a super smart AGI but humans assigning too much "intelligence" (and anthropomorphised sentience) to the next generations of LLMs etc and thinking they are way more capable than they actually are.

thehalf13, 11 months ago

While I see where all of the concern comes from, there are major problems when it comes to trying to regulate current technology. The first problem is that many of the people writing legislation don't understand how the technology works or what their legislation is actually saying in terms of the technology. I will use a US federal judge as an example where he ruled that all ISP's and other providers had to make it so that this one website can never show up again(oversimplification for brevity). This is literally impossible because you can't ban all domains they might possibly get and even if you could ban the domains, you can't ban them from showing up as a different entity with a different IP.

The next major problem that I see particularly in the US is that technology is changing faster than we can understand it as a species. We have entered into an era where the technological landscape changes every 10 years, and we can't even fully figured out equality rights despite having this fight for more than 50 years.

The next problem ass I see it you run into with technology regulation is the regulations need to outpace the technology. One you let it out you can't really put it back. Let's take some of the AI's currently. Creators found out their AI was being used for things they didn't like so they removed that capability. This just led to the people that were doing that to creating another instance of an AI that would.

And the last that I'm going to bring up is that like you mentioned it has to be a global effort. Internet and any technology that uses IP base anything is a global technology. Oh we want to limit this to this one geographic area, that's fine I'll just fire up a VPN. Oh you blocked all the major VPN IP addresses, that's cool to I'll just get this cloud server in the country I want with a static public IP and VPN tunnel into it and now I'm still in that country. Unless they literally block off networking with other countries, which causes major major issues, there is no way to stop that.

I would love to see some good regulation happen that ensures everybody is safe from the dangers that technology brings. But I guess that I'm a pessimist by thinking that it may be to late.