On the one hand, I agree with that observation. On the other hand, it makes me think of the old (half a century or so) argument "but that is not AI, it is just X. It would have to do Y to be real AI", where this year's Y will, in 5 years time, be solved by a new X.
Yes, those surgeons in some sense are just mechanics/craftsmen, but what they do was research five years ago and they know they must stay up to date, so that they can do that Y by the time that new X has sufficiently developed. Also, many of them work on getting those new Xs to market, if only by helping in trials.