This is a very popular trope among people, especially since CGP Grey's "Humans Need Not Apply" video, but it just isn't the case that automation will drastically increase long-run unemployment just because certain jobs will disappear.

• David Autor's Polanyi’s Paradox and the Shape of Employment GrowthAutor is notable here has he has massively advanced our understanding of the interaction between technology & labour over the last couple of decades, he posits automation as an extension of the Skill-Biased Technological Change hypothesis which represents manageable inequality changes (this is wage inequality, labour/capital shares remain stable but there is a clear divergence between types of labour actors) but no structural employment issues. The absence of structural employment is expected based on the way we understand technology to act on labour, as a productivity multiplier, and even if the SBTC hypothesis turns out to be incorrect this does not imply structural employment but rather a different form of inequality.
• Daron Acemoglu's Why Do New Technologies Complement Skills? A more comprehensive discussion of the SBTC effect.
• The Future of Unemployment: How susceptible are jobs to computerisation? The oft-cited paper and the first to do a thorough review of the scale of labour disruption that may occur in the future.
• Robots Are Us: Some Economics of Human Replacement Built on the prior paper to examine some of the productivity effects in a simple tech & non-tech worker model. While some inputs to their model do produce a result which suggests a reduction in labor demand their conclusion is that the likely outcome is the other kind of inequality (declining labor share) but again with a clear policy solution, they also have alternative policy solutions for avoiding the unlikely scenario of net labor demand falling.

More generally it's argued that, historically, automation has not reduced employment. Automation has historically acted as a multiplier on productivity which drives demand for human labour. Pre-singularity its very hard to imagine this changing, we will undoubtedly encounter disruption effects (people will have the wrong skills, their earnings will reflect this matching issue rather than unemployment doing so) but from an economics perspective there is little difference between replacing a field worker with a tractor and an office worker with an algorithm. Certainly the office worker needs to find a new job, if they don't have demanded skills that job may not offer earnings growth opportunities but it doesn't imply unemployment any more than the mechanization of agriculture did.

DEE JAY KHALED

SAY MY NAME BABY