Harmless prompt blocked for safety, still counted against quota

As the title says, this Veo2 prompt (with slight variations):

“camera panning horizontally on the fronts of a series of old shops, in the old neighborhood of a big European city”

has been blocked three times for “safety”. First of all, how is it “unsafe”? And more importantly, why has the quota been consumed the same? How can I consume a compute quota if no compute happened (other than the safety filter, so not much at all)?

I do realize we’re all figuring this out but come on…

@unmarco,

Welcome to AI developers forum, Thank you for flagging this.

i have tried the prompt in AI studio and seem to get the result

as you pointed we’re continuously improving the models every day and your feedback is valuable.

could you please let us know if you were using any config parameters and were you using sdk or AI studio so that it can be helpful zeroing on this issue