David Autor

They're highly secure.

They're very expensive.

And we are dependent upon them for health care, for our children's education, for our legal services, for the code that they write for us, for the buildings they build.

And so we have a collective interest

in making those things more productive, more affordable, but also ideally having more people engaged in the middle of that who can do that work and where it creates opportunity.

I think the fundamentally valuable human skills are going to remain, you know, one, specialized domain knowledge in some area where they can.

And two is really the sort of judgment and flexibility to work in a very noisy environment where you have all kinds of inputs, all kinds of data of unknown quality.

And you have to make high stakes decisions to do that.

Well, people need to be able to communicate.

They need to be able to think analytically.

They need to actually understand statistics, sadly, and make sound decisions in this very complicated world.

AI offers great tools for that.

It also offers great risk because people can use it in a way that interferes and undermines the learning process.

And that, I think, is a totally separate discussion, but a very legitimate concern.

You should recognize, even more in the labor market, AI has great potential and great risk.

Productivity growth is not the problem that most people should be worrying about.

This was also the case in the 1960s during the Kennedy administration, the Johnson administration, the Blue Ribbon Commission on Automation and says, if we keep having this much productivity growth, there just won't be enough demand for us to use up all this stuff.

And people run out of work because we're just making too much of everything.

This has never occurred.

People's consumption desires rise like at least as fast as productivity growth.