The creativity gap

I really enjoyed this article. I understood it from the more general lens of artificial creativity, or the creativity gap in AI.

If search and learning comprise the two main modes of AI, creativity represents an unexplained third way. For search and learning, the assessment criteria is known: these axes may be explored and rewarded. But for something that’s genuinely creative, the assessment criteria is unknown. In science, this gap requires a conjectural leap; in mathematics, intuition; in art, it’s the novelty that separates the outstanding artists from the mere imitators.

Your examples also make the value of imitation quite concrete. The results are often superior to human imitators. But based on the mechanisms of these systems, we know they’re not genuinely creative.

As you point out, the main challenge in the creative realm is the very thing that defines creativity: the assessment criteria doesn’t exist until it’s created (at least not explicitly within the system). It suggests we need a new way to think about the problem, a philosophical advance, to make substantial progress.

Thanks for your wonderful article!