5 Comments
User's avatar
H Moussawi's avatar

Very interesting findings, and consistent with the anecdotal evidence from the news media...

Expand full comment
A11's avatar

Damn it, I don't wanna have to switch from CS to nursing.

Expand full comment
Alex Zarifeh's avatar

Brilliant work Bharat, thank you & to your co-authors!!!

Expand full comment
Jordan Peeples, PhD's avatar

Awesome work and great read!

So, an interesting fact is that youth unemployment has grown relative to overall employment since the Great Recession. During Covid, that relationship naturally decreased a bit since unemployment shot up for all groups. I wonder what the pre-Covid trends here look like for these now AI-displaced jobs in your paper? We may see the same results, but I think it may be an important factor to consider to ensure this isn't structural.

I also know that job postings for software engineers far surpassed all other job descriptions (from Indeed) in 2022. This could also be an equilibrium issue, but perhaps it's addressed in the empirical design in the paper.

Just thought I would share these! I think my first point addresses the recent youth issue as it's not quite so recent.

Expand full comment
Carsten Bergenholtz's avatar

Excellent post. Concerning "Why entry-level workers?": I wonder if AIs (perceived) impact on the job search process also plays a role. a) I could imagine that some companies are worried about hiring new graduates, since they might worry that grades or diplomas don't really showcase what they can actually do, in a world where AI might have helped them get the diploma. b) Relatedly, given how much easier it is to write (a lot of) job applications that are tailored to the particular position/company, the ones hiring might fint it more difficult to figure out who to hire - maybe "informal networks" are getting more important, and of course new graduates have fewer of those.

Expand full comment