AI developers are arguably among the most powerful agents in the modern marketplace today. While they central to the business and science of AI, they are largely marginalized as ethical agents. If they are invoked at all in public discourse, it is most often for their status as “tech bros” – mostly male, mostly white, mostly Western, but almost always as insufferably overconfident know-it-alls. The academic literature, likewise, is predominantly concerned with how to produce more ethical algorithms, not necessarily more ethical AI practitioners. This talk will move the algorithms and the companies deploying them into the periphery. It will instead center the ethical agency of AI developers. I will discuss findings from interviews with developers about how they conceive of their ethical agency, the dilemmas they face, and the excuses they use to avoid taking responsibility for their creations. I will also show how those seeking to influence developer ethics (the authors of AI ethics guidelines) likewise hide the ethical agency of AI developers through linguistic and semantic practices of backgrounding and suppression. The talk will conclude with some recommendations for how to move forward, including how to engage AI developers in their own ethical development.