Modeling AI-driven occupational change over the next 10 years and beyond

Hello everyone! Occasional lurker, first time poster here (working in government tech ops, writer in my free time). I learned a lot from this forum when I wrote an exploratory piece about 6 key AI variables to watch last October.

Since then, I’ve taken a deeper dive into how quickly occupations typically change each decade (going all the way back to 1870!) to get a better sense of what a reasonable timescale for AI-driven automation of labor looks like. There was an interesting debate between Ajeya Cotra and Jason Crawford a few weeks ago about whether this would take years or decades.

I take Jason’s position on this (decades), but that still leaves me predicting change much more quickly than the official U.S. government projections, which I’ve found to be extremely conservative, especially given predictions like Leopold Aschenbrenner’s re: AGI in 2027.

My most recent piece focuses on three occupations I think might actually grow faster even if we see powerful AI emerge in the next decade. I see these ones as archetypes of the three types of work I could see continuing to exist for some time to come:

  • Jobs where people are needed to be personally accountable for factual accuracy in fields where there will continue to be emerging knowledge (e.g. biologists, legal experts, etc.)

  • Jobs where having a human is inherently valuable to the customer (therapists/​counselors, and personal service workers among others)

  • Jobs that require significant management decisions to be made by a directly accountable individual (managers, whether directly by regulation or downstream of it, or entrepreneurs)

For now, I’ve left out “jobs that require physical work, specific expertise and can perform in unpredictable environments” as I see this as a bit of a question mark over the very long run. But I’m interested in doing a more specific breakdown of this sort of work in the future.

This list is perhaps a broader one than the one that @Roger Dearnaley proposed a few months ago, but I’m interested in further modeling this out in a way that hopefully helps improve official estimates and government readiness for the possible job market impacts of AGI (e.g. in these three occupations). I’d love to connect with anyone else who is working in this space, or is aware of any funding available for this type of research.

I’d also appreciate any feedback on how I’ve broadly categorized these AGI-proof jobs, or about another hypothesis that I have: that “drop-in remote workers” could be a bigger threat to new workers than experienced ones; organizational dynamics might mean the complete lack of hiring could have a bigger effect on employment, at least for the first few years (given rising youth unemployment rates, we may already be seeing this dynamic at play).

No comments.