Dear Jeff, string theory has had a colossal influence on the renewal of enumerative geometry, a two century old branch of algebraic geometry inextricably linked to intersection theory.
Here is a telling anecdote.

Ellingsrud and Strömme, two renowned specialists in Hilbert Scheme theory, had calculated the number of rational cubic curves on a general quintic threefold by arguments based on their paper 

*On the Chow ring of a geometric quotient*, Annals of Math. 130 (1987) 159–187

Their result differed from that predicted by string theory. Of course everybody thought the mathematicians were right, but actually there had been a programming error in their calculations and the correct result was that of the physicists (which Ellingsrud and Strömme confirmed after fixing their bug). 

This was the beginning of a long list of results predicted by string theorists and subsequently proved by mathematicians, a celebrated example being Kontsevich's formula for the number $N_d$ of degree $d$ rational curves in $\mathbb P^2$ passing through $3(d-1)$ points in general position. 

You can read all about Kontsevich's formula in Kock and Vainsencher's fine book: http://www.mat.ufmg.br/~israel/jojoEE.pdf