>We have AI to raise flags

That it would like get wrong more often than not. No thanks.