The AI field needs translational Ethical AI research

Calls for Ethical AI have become urgent and pervasive, especially as ethical issues surrounding products at tech companies are increasingly scrutinized by the public. Yet even after a first wave of responses to these calls coalesced around principles guide decision-making second generated technical tools mitigate specific issues, multiple lines evidence indicate that only limited impact on daily practices users producers. In other words, there is big gap between what we publish in academic papers creators need generate reflect society's values. no means field this problem. However, when medical ecology fields documented similar gaps their fields’ scientific discoveries people actually use, they invested tremendous resources into subfields developed about how translate was done lab adopted solutions. I argue commentary it our research community's moral duty invest own subfield “Translational AI” will determine best ensure practitioners can implement venues production settings. Further, offer concrete steps doing that, drawing insights gleaned from translational fields. Closing “Ethical Publication-to-Practice gap” be considerable transdisciplinary challenge, but one community has unique expertise, political leverage, responsibility tackle.