An interesting test would be if we had, say, 50 good AI generated paintings and 50 good real artist created paintings mixed in together randomly, and your job was to identify which was AI and which was real.* I suspect for a more average art critic (which would be far above me!) it would be...