Designing AI, Projecting Democracy

As concerns about the impact of AI on contemporary polities proliferate, so do ways to address the democratic deficit of these technologies. Increasingly, the issues of bias, accountability, transparency, and discrimination come to the forefront of public discussions on AI, highlighting the fundamentally political nature of these technologies. In an effort to align the development, implementation, and scaling of AI to public values, scholars, policy-makers, and activists try to envision how these technologies can be democratized.

This way of designing AI with democracy emphasizes the separation between technology and politics. From this perspective, AI and democracy are confined entities, rather than dynamic and interdependent ones. To think of AI for democracy, rather than with it, entails seeing technology as actively partaking in the constitution of political life as such. Seen this way, both AI and democracy become imbued with the dynamism that brings questions of what polities, in which they are present, we see as ethically and politically desirable. Taking a cue from political philosopher Arendt, plurality is an important condition of democratic societies, whereby plurality and political activity “[are] valued not because [they] may lead to agreement …, but because [they] enable each citizen to exercise his or her powers of agency, to develop the capacities for judgment.” Democracy here means finding a way of being together amid multiple perspectives and plurality is about crafting a space for public interaction and deliberation.

This way of thinking means that both AI and democracy are fundamentally open endeavors. Instead of seeing the democratic project as already ‘finished,’ ‘given,’ and something that needs to be protected from technology, AI for democracy underscores the changing and dynamic relationship that these entities mutually form.