@dingemansemark and unfortunately, the evidence I’m collecting suggests that this risk of being flagged as “AI” is higher for L2 writers and neurodivergent ones (I’m in the latter category but not the former) even when it’s other humans, not automated tools, doing the “detection”.