This is a hot topic I've heard a lot on podcasts recently. I've heard it in podcasts in the Financial, Tech and Miscellaneous categories. The common trend is they think it's coming sooner or later. The trend isn't new. My photographer friends and a typesetter friend lost their jobs due to new technology making their skills obsolete years ago. Perhaps the trend is increasing exponentially? I know they're trying to simplify the management of computer networks (my field), but their success here is limited so far.
Arguably your photographer friends could have transitioned to digital technologies, Photoshop, online services, etc. This is the usual counter-argument, and often people will throw in a "Luddite" aspersion as if that settles the matter. More serious rebuttals will be a bit more specific: But then there are counter arguments: And some questions: This is extracted from: http://philosophicaldisquisitions.blogspot.com.au/2014/08/are-we-heading-for-technological.html
I'm sure this will be discussed a lot but I just wanted to throw in one thought I had along these lines. I think that the last industry that will be taken over by machines is entertainment. If machines can actually be creative enough to actually provide us with satisfying entertainment then they will be for all intents and purposes human themselves. And if you look at the growth of the entertainment industry through all media for the last say 100 years, it is probably exponential with more people working in it now and more output than ever before.
Imagine some hypothetical software for a future iPhone or PC that could create all the music you wanted according to your particular preferences and feedback (feedback via the increasingly ubiquitous and sophisticated bio-monitoring gadgets that maybe one day could gage mood and physiological response to different compositions at different times for different activities.). That could destroy the music industry and they wouldn't go down without a fight. Movie/TV creation might be far more difficult.
I'm a programmer and it's been interesting to see the evolution of my field in the last decade or so. When I was started there was much more low-level work to be done, getting things to work, writing database access code, that kind of thing. Today, much of that is machine generated now and my work tends to be much higher level. There has been a huge boost in productivity also, things just get done much faster now and usually to much higher quality.