Nationalize AI - Part 1 - Defining The Problem

This is Part 1 in a multipart series on nationalizing AI:

• Nationalize AI - Part 1 - Defining The Problem
• Nationalize AI - Part 2 - Labor
• Nationalize AI - Part 3 - Corporate Power
• Nationalize AI - Part 4 - Inefficient
• Nationalize AI - Part 5 - Superiority
• Nationalize AI - Part 6 - Conclusion

As we’ve seen multiple breakthroughs in AI - I have been thinking a lot about the short, medium, and long term consequences. I have come to the opinion that, over some undefined but likely relatively short timeline, there are really two potential futures.

I do have a predisposition towards the nationalization of certain industries, particularly ones that are extremely harmful because they are run privately (health insurance is one such example, but hardly the only). While each of these industries becoming publicized would represent large quality of life improvements for the American people - AI nationalization dwarfs all the others by multiple times.

What do I mean by consequences? Well, I’d like to frame the discussion in these areas of concern.

• Effect on labor
• Corporate power
• Inefficiency of private AI
• AI Superiority

Now, obviously - each of these are extremely interdependent and so it is a bit hard to cleanly divide the supporting points of each section; but I will do my best.

Assumptions

The main assumption I’d like to put forth is that AI is powerful and will continue to accelerate. The time frame is in question, but the assumption is at some point we will have AI more intelligent and capable than any human or group of humans.

What could nationalization look like?

Well, at a high level - in my scenario, it could go something like this:

1. The federal government creates its own AI lab
   • The law must define the goals of this lab and regulations therein.
   • It would be critical that it is a program with transparency, oversight, and regulation more comprehensive than any existing program.
2. It outlaws private research on AI and/or direct ownership of certain / amount of GPUs/chips.
   • You’d also need to outlaw citizens working on AI in other countries.
   • This may feel radical - however there is precedent for such an action. Consider that any given private company or citizen can not develop a nuclear weapon.
3. Consolidate the existing AI labs under the federal government’s AI lab
   • The goals and research direction would likely change from the current private labs, but its goals and research direction would be clearly defined and transparent to the public.
4. Institute direct rebate programs to citizens (especially to anyone affected directly by job loss due to AI)
   • Perhaps direct rebate is not the correct way to go here, but the point would be to socialize the benefits of AI - however that looks.

My thesis, which I hope to convince you of by the end of the series, is that we must nationalize AI; and to a lesser extent, that it should look like the outline above.

In the next part of the series, we will discuss the effect of AI on the labor market and why it demonstrates a strong reason to nationalize: Nationalize AI - Part 2 - Labor