As AI displaces jobs, the US government should create new jobs building affordable housing

What if some displaced workers help build the housing we need?

We have a housing shortage in the U.S., and it is arguably a major cause of long-term unrest about the economy.

Putting aside whether AI will eliminate jobs on net, it will certainly displace a lot of them. And the displaced people are unlikely to be the same people who will secure the higher-tech jobs that get created. For example, are most displaced truck drivers going to get jobs in new industries that require a lot of education?

Put these two problems together and maybe there is a solution hiding in plain sight: create millions of new jobs in housing. Someone has to build all the affordable homes we need, so why not subsidize jobs and training for those displaced by AI? These jobs will arguably offer an easier onramp and are sorely needed now (and likely for the next couple of decades as we chip away at this housing shortage).

Granted, labor may not be the primary bottleneck in the housing shortage, but it is certainly a factor and one that is seemingly being overlooked. There are many bills in Congress aimed at increasing housing supply through new financing and relaxed regulatory frameworks. A program like this would help complete the package.

None of this has been happening via market forces alone, so the government would therefore need to create a new program at a large scale, like the Works Progress Administration (WPA) at the end of the Great Depression, but this time squarely focused on affordable housing (and otherwise narrowly tailored to avoid inefficiencies).

There are a lot of ways such a program could work (or not work), including ways to maximize the long-term public benefit (and minimize its long-term public cost), but this post is just about floating the high-level idea. So there you have it. I’ll leave you though with a few more specific thought starters:

• Every state could benefit since every state has affordable housing issues. Programs become more politically viable when more states benefit from them.

• Such a program could be narrowly tailored, squarely focused on affordable housing (as mentioned above), but also keeping the jobs time-limited (the whole program could be time-limited and tied to overall housing stock), and keeping the wages slightly below local market rates (to complement rather than compete with private construction).

• It could also be tailored to those just affected by AI, but that doesn’t seem like the right approach to me. The AI job market impact timeline is unclear, but we can nevertheless start an affordable-housing jobs program now that we need today, which can also serve as a partial backstop for AI-job fallout tomorrow. It seems fine to me if some workers who join aren't directly displaced by AI, since the program still creates net new jobs we will need anyway and to some extent jobs within an education band are fungible.

• We will surely need other programs as well to help displaced workers specifically (for example, increased unemployment benefits).