this "experiment" notwithstanding, "what does it mean if AI can do X better than we can?" is STILL an irrelevant question—the best AI can philosophize EXACTLY as well as we can bc it's a machine that spits out the probable average of the human philosophy in its dataset.
https://www.wired.com/story/an-acclaimed-book-about-digital-manipulation-was-actually-written-by-ai/