I disagree - I think it's because there's likely insufficient membership control to have a worthwhile signal-to-noise ratio. AI discussion is notorious for this, because of the quantity of cranks around.

Furthermore, legitimate AI researchers know that there are many cranks around, so they're less interested in starting under-moderated and under-restricted discussion. So when someone comes around asking to start under-moderated and low-barrier-to-entry discussion on AI, it signals that they're likely a crank.