AI for Good vs. AI made evil: why ethical AI should be a top priority for insurers