I don't see how their conclusion follows. The way I see it: - They gave people two randomly generated maps - They told the people that the second map was AI

I don't see how their conclusion follows. The way I see it: - They gave people two randomly generated maps - They told the people that the second map was AI generated - There were differences between the maps (because, well, random) - People noticed these ("no one found it equally challenging") - People attributed the differences to the AI, because that's what they were told This would be a valid conclusion to make if the players had been told that the AI did something specific and there were consistent claims that said specific thing happened. This, on the other hand, is just a study of how randomness works.