These are wolf cubs the villagers are raising themselves and proudly showing off at wolf cub fairs (so cuuute!), expecting them to be great help with the actual shepherding once they grow up. The popular concern is future shepherd unemployment.
(Imagine crying "wolf" in this narrative.)
Wrapping up arguments 1 and 3 against into one, if we cry wolf at a chihuahua, or cry wolf for any reason but that we think there is a wolf, we are using language in a way bumps our own discourse level up a simulacra level or two. So, we are using language not to communicate our best estimate about reality but to communicate group allegiance or suggested courses of action.
As you say, one consequence of this is that the metaphorical villagers may justifiably decide to trust us less, because they realized we are not using words to describe reality. The more important consequence to me is that our own ability to communicate, coordinate, and describe reality will become impaired.
I think just truthfully crying "let's prepare for real wolf", because they clearly didn't train the miniature one well enough, has fewer downsides.
I think the parable/metaphor is hurting much more than helping. Not only is the fable pretty stretched as to actual villager behavior, the threat and context are FAR removed from any actual modern behaviors.
Propose an actual action, and analyze it's likely effects.
While you can "cry wolf" in maybe useful ways, you can also state your detailed understanding of each specific situation as it arises and how it specifically plays into the broader AI risk context.
Epistemic status: I am confused and trying to become less confused.
So, our shepherd community noticed the wolf. It's a small wolf, about the size of a chihuahua, and is unlikely to cause any serious damage. Should we cry "wolf"?
What is the relative weights of these arguments?
What arguments did I miss?