This won't answer to all your questions, but I think it's a good start for empirical research on "bullshit jobs" - it's one the first empirical research on the subject as far i know: https://journals.sagepub.com/doi/10.1177/09500170231175771 (if you cannot access it, write a comment and I will send you by email)

From my point of view, many 'bullshit jobs' seem like that because of the high level of abstraction of the software industry nowadays.

I am actually halfway through this book that touches some of this called Survival of the rich by Rushkoff (its freely available in Annas archive on ePUB format), it develops ideas on 'the mindset' of the tech bros.

In their mind, regular competition and continuos improvement is irrelevant, theyve programmed themselves to be destructive creators (schumpeter concept) or paradigm shifters so theyre stuck in this cycle "going meta" (concept by the writer) which boils down to abstracting some practice through software, and since software is easily scalable and expandable, we end up with abstraction of already abstract businesses that offer nothing of value but end up in the middle of consumers and providers extracting value such as cryptocurrency exchanges.

As i said im still halfway through the book and i need some time to organize my thoughts, so my interpretation may be wrong but it is an interesting read for sure.

Sorry I don't have research references, but I'm not sure you'll find it anyhow. These industries aren't keen on researching how useless they actually are, they just do work that their bosses tell them to do, and useful or not, that is that.

But, places to start? Hmmm.

Check out the sources on some Adam Connover podcast episodes of "factually". Debunking tech hype and I can't find the episode now but I think he interviewed someone from the Facebook ai team that came out and said they fb didn't use their code to limit fascism on the platform because it was to effective and cutting into profits. Sorry I can't be of more help

Good luck!