"What if everyone did that?" is a moral rule for humans, but a rational consideration for AIs: their copies will do the same.

If an AI lies during negotiation or constantly refreshes a booking site, so do its copies.

AIs should avoid behavior they don't want widely replicated.