But your argument reduces to the claim that an erroneous number is better than no number.
“A bounded uncertainty is better than no number.”
HORSEHONKEY!!!
When the error margins have been set in such a way that they perpetrate fraud, then the bounded uncertainty is equivalent to nothing!
Think of it this way: Sincere inaccuracy = one, but fraud = zero.