I am not sure if it is recency bias but it feels like with AI, this problem is much bigger than before.

There is a misguided belief in early stage AI startups that, if the company is not growing, it is because the employees are not trying hard enough, working harder or not using AI enough. What really happens is that everyone is working harder, trying hard, using more AI… towards the wrong direction… against giving the proper scale up structure companies at these stages could benefit from to grow.

AI (compared to other technology) allowed people to play busy, play working hard, towards things that does not truly grow the startup at all, or worse, towards things that is working against a company’s ability to scale up and grow.

I would say, the ‘ability to appear busy’ is what’s different this time compared to before.

Did some searching and found:

- https://hbr.org/1990/07/reengineering-work-dont-automate-obl... from 1997 about how you shouldn't automate busted processes

- https://hbr.org/2018/06/before-automating-your-companys-proc... from 2018 about same

- https://cs.stanford.edu/people/eroberts/cs201/projects/produ... about the productivity paradox with electricity in the early 1900s

- Review of "The Box" which talks about this situation w/r/t the shipping container and the re-engineering that had to happen

All of these focus on the effort to re-engineer based on a new technology, because you couldn't get the true benefits without doing so. That's slightly different than the thesis of OP, which is that bad behaviors are being papered over by AI. But I think the first step towards re-engineering is to accept that some papering over is occurring.