Jacob Miller examines how AI could drive long-term productivity gains in old, stagnant industries
Transcript
Where is productivity growth likely to come from? The topic of the day, and likely to be the topic of the day for some time to come, is artificial intelligence. There’s a lot of buzz, a lot of hype, probably some buckets of froth, but also real productivity enhancements that could come about across largely traditional sectors over the next couple of decades.
We want to actually understand the impact of AI, not just believe the hype. And I have a simple framework for trying to understand how this is going to flow through. We can look across sectors, understand what’s the weight of that sector, how fast we expect that adoption to happen in that sector, what’s the long-run potential for AI to disrupt or improve this industry, how high quality are AI models at the problems in this space, and how elastic are the outputs to being supplemented with AI.
Two sectors I’m most excited about are health care and real estate. These have had very little improvements in productivity over the last 40 years and are the largest single parts of the average consumer’s spending. Health care costs have ballooned. Rental costs and home buying costs are up, and building and construction is extremely laborious. We are pretty excited about how these more traditional industries can be transformed with boring AI.
Adoption is likely to be relatively slow. These are not fast-moving spaces, but the models seem very attuned to the work in health care—being able to do advanced diagnoses, pre-populate patient charts, do hospital billing and servicing, back office. Models are already able to do that today, and they’re improving every year.
On the construction side, actually building the building may be less AI-focused, but permitting, expense management, procurement, and planning are things that are very well suited to AI. So while it’s not the most elastic sector, the models are relatively high quality, and the potential lift, given the costs and time involved in those processes, is pretty extreme in the sector. It actually seems like a great space for AI to disrupt.
If we can change just these two industries up to their potential, that’s going to be a material change to how people live their lives in America every day, and a potential massive shift in productivity.