Computing will continue to aid mathematical discovery, proving (useful) theorems like Four Color, or even formulating their own (useful) theorems before proving them.

For example, in Einstein's Special Theory, there's not a lot of deep mathematics (relative to the General Theory) so it's not too difficult to imagine Watson taking the problem and axiomatizing (just to see what would happen) that light speed is always constant (which is the key to everything else that followed).

But what about Descarte's insight that geometry could be described by algebra using a coordinate system, giving birth to analytic geometry? Or proving that it can't prove all true statements in arithmetic (Godel)? Or even asking the question?

I think we're a long way off from doing something like that in a machine.