The ethico-politics of design toolkits: responsible AI tools, from big tech guidelines to feminist ideation cards

Abstract

Abstract This paper interrogates the belief in toolkitting as a method for translating AI ethics theory into practice and assesses the toolkit paradigm’s effect on the understanding of ethics in AI research and AI-related policy. Drawing on a meta-review of existing ‘toolkit-scoping’ work, I demonstrate that most toolkits embody a reductionist conception of ethics and that, because of this, their capacity for facilitating change is limited. Then, I analyze the features of several ‘alternative’ toolkits–informed by feminist theory, posthumanism, and critical design–whose creators recognize that ethics cannot become a box-ticking exercise for engineers, while the ethical should not be dissociated from the political . This analysis then serves to provide suggestions for future toolkit creators and users on how to meaningfully adopt the toolkit format in AI ethics work without overselling its transformative potential: how different stakeholders can draw on the myriad of tools to achieve socially desirable results but reject the oversimplification of ethical practice that many toolkits embody.