There's certainly a risk that some kinds of abuse patterns will get through our reply-to-publish model. Eventually we'd like to try and detect bot traffic and challenge them with a CAPTCHA or similar.