I'm not super familiar with the inner workings of the IMF, but it does state in the notes section state that these "discussion notes" don't reflect the opinions of the IMF, just the authors.

In addition, reports like this are always tricky. It seems to base their predictions on where people think the technology could go, rather than where it is today. For example they have professionals as the highest to be potentially impacted. This shows they are speaking about generative AI (as previous AI/ML didn't generally touch these fields). While generative AI has been impressive, I'm not sure anyome has been able to demonstrate an substantive impact to professionals.

A counterpoint though would be something like customer service. Automation in that space over the past couple of decades has reduced the cost/number of workers, but at the cost of efficiency/efficacy. So it's possible we may see things like AI doctors and lawyers, but potentially at a reduced quality for the average person.

Just my thoughts regarding the angle this paper/discussion notes went in.