Locating the work of artificial intelligence ethics

The scale and complexity of the data and algorithms used in artificial intelligence (AI)-based systems present significant challenges for anticipating their ethical, legal, and policy implications. Given these challenges, who does the work of AI ethics, and how do they do it? This study reports findings from interviews with 26 stakeholders in AI research, law, and policy. The primary themes are that the work of AI ethics is structured by personal values and professional commitments, and that it involves situated meaning-making through data and algorithms. Given the stakes involved, it is not enough to simply satisfy that AI will not behave unethically; rather, the work of AI ethics needs to be incentivized.