Yes, but those use cases either doesn't matter or isn't unattended. A lot of tasks aren't very important and amenable to high error rates with human oversight. I don't think the AI architecture that will cause the expected mass technological unemployment currently exists.

I don't expect the good case in the near term. Perhaps after 2050 when attrition has claimed the bulk of the 60s generation, they will be the driving force against the good case.

>The Economics of Automation: What Does Our Machine Future Look Like?