I think what he's doing would be more acurately described as Machine Learning, rather than AI. His algorithms are producing better logo proposals based on the data it is collecting from other users, which fits a common definition of ML [1]
[1] A computer program is said to learn from experience E with respect to some class of tasks T and performance measure P, if its performance at tasks in T, as measured by P, improves with experience E (Tom M. Mitchell, 1997)
Machine learning is a better term because it's more specific (and less controversial). But under most definitions of AI, ML is a subfield of AI, so calling it AI may be suboptimal, but not wrong.
Personally I'm a fan of the AI definition "things humans can do and computers can't do yet", but I recognize that that definition isn't terribly useful.
[1] A computer program is said to learn from experience E with respect to some class of tasks T and performance measure P, if its performance at tasks in T, as measured by P, improves with experience E (Tom M. Mitchell, 1997)