Until we have adequately responsible AI, here's how we can protect ourselves from opaque and thus dangerous AI already affecting our lives.