AI is changing the world and nobody knows what is going to happen.

Will we see an intelligence explosion or have a slow takeoff? Will AI lead to a new era of human flourishing, to autocracy, or to something much more mundane? How much work will get automated and how quickly?

All these things are subject to wide uncertainty. Smart people have widely differing opinions.

So, dear reader, what (if anything) should YOU do given that there is so much uncertainty?

I think there are a few no-regrets actions you can take if you - like me - you take the potential of AI seriously. Some actions that will be beneficial under a wide range of possible outcomes (1).

There are a few basic assumptions that underly my thinking:

AI will transform society. This has happened before with technologies like agriculture, the steam engine, and the computer. It is already starting with AI.
Nobody knows how the transformation will go. There is just a lot of uncertainty about what AI will be able to do, how well we will be able to control it, how it will be deployed, and on what timelines.
It might be bad. AI could enable terrorists to create powerful biological weapons. It could lead to mass unemployment. We could - wittingly or not - give AI systems access to powerful weapons that they decide to deploy.
AI will take over a lot of work. AIs already do an OK job of writing copy, illustrating, writing code, and doing research. They will only get better.
The AI transformation will occur under circumstances of competition. Both between nations as well as between companies.

Those are my starting points. I may write a future post on how you might think differently if you disagree with one or more of these starting points.

So now, let’s turn to the first thing that comes to mind in a situation like this. What’s going to happen to me?

How do I ensure the AI transformation goes well for myself?

If AI will take over a lot of work, what should you do to make sure that you will be ok if AI takes over a bunch of jobs?

There are four areas to focus here:

Use the AI. Become great at getting things done faster and better using AI. This blog post by Ethan Mollick has some great suggestions on how to get good. They boil down to “just use it a bunch and figure out what it’s good at.”
Know people and be likable. As more and more of the “thinking work” gets done by machines, a lot of the “people work” will still be left to humans. Your network will become more and more important. So be someone people know and like.
Manage. Know how to manage people well. And if relevant for your job, know how to manage AIs too.
Save more. To have a cushion against future shocks.

Now that your career is future-proof, and you have ample savings in the bank, let’s turn to what you can do to help the rest of us out.

How do I ensure the AI transformation goes well for the world?

Everyone should have at least one political cause of choice. If you don’t have one already, it is a great time to make AI your political cause of choice. Especially if you live in the US or China, since these are the two leading countries in the development of AI, and are both reasonably responsive to the concerns of their people.

But! Uncertainty. Is it better to Nationalize parts of AI development, or leave it largely in the hands of private companies? What kinds of regulations should be placed on AI development and deployment? To what extent should Americans be focused on getting the best models before China at all costs?

There are a couple principles that seem to make sense under a broad range of paths that AI development could take:

Ensure transparency about what developers are training AI models to do. About what values are being instilled in the models. About what kind of safety measures are being programmed into them. The more these models influence the world, the more critical it is that they are subject to scrutiny by the widest possible number of stakeholders.
Ensure adequate cybersecurity to ensure dangerous AI tools don’t get into the wrong hands.

Make your voice heard! The development of top AI models should happen transparently and securely.

It also seems valuable, even without the pressure of AI, to ensure US-China competition doesn’t turn violent. This is an area I am trying to learn much more about. If anyone has any concrete ideas on how we can ensure we don’t have war, let me know ;) (2)

Closing thoughts

AI is changing the world and nobody knows what is going to happen. The above is my best guess - right now - for how you can give yourself and the world the best chances of success during the AI transformation.

What do you think?

—

None of my thinking in this post is original. And this is very much a work in progress capturing my thinking at this point in time. I would love for you, dear reader, to tell me what you think. Also note that this advice - especially the political advice - will make the most sense for white collar Americans.
My vague thinking here is: It’s inevitable that the US and China will compete. This is what global powers do. But the more we can channel our competition into productive ends, the better. For example, you can imagine countries focusing their competition on achieving soft power dominance by providing paths for economic development, winning at sports, and exporting great culture rather than fighting wars.