Artificial intelligence (AI), particularly various methods of machine learning (ML), has achieved landmark advances over the past few years in applications as diverse as playing complex games, language processing, speech recognition and synthesis, image identification, and facial recognition. These breakthroughs have brought a surge of popular, journalistic, and policy attention to the field, including both excitement about anticipated advances and the benefits they promise, and concern about societal impacts and risks – potentially arising through whatever combination of accident, malicious or reckless use, or just social and political disruption from the scale and rapidity of change.