Setting Democratic Ground Rules for AI

by Beth Kerley, Senior Program Officer, International Forum for Democratic Studies

This Big Story was adapted from Beth Kerley’s recently published report for the International Forum for Democratic Studies, Setting Democratic Ground Rules for AI: Civil Society Strategies

Advances in artificial intelligence (AI) are changing the playing field for democracy. Since social media’s emergence as a tool of protest, commentators have regularly stressed how our evolving technological landscape is transforming our political world. The digital tools on which we rely help to determine how people express themselves, find like-minded communities, and initiate collective action. As the International Forum has tracked in our “Making Tech Transparent” series examining AI surveillance, smart cities, and the digitalization of governance, these technologies also affect how governments monitor people, administer services, and—in some cases—dole out repression.

We are poised for another seismic shift in the balance of power between people and governments. With recent leaps in the development of large language models (LLMs), the global proliferation of AI surveillance tools, and growing enthusiasm for the automation of governance processes, technological advances are already challenging democratic systems. Recognizing that digital advances themselves do not necessarily work to democracy’s benefit, prodemocratic stakeholders must work to erect guardrails around AI development and deployment; ensure consultation with communities whose democratic rights may be impacted by AI systems; and chart development trajectories that infuse AI technologies with democratic values. The consolidation in dictatorships of authoritarian models for the integration of AI—which reject privacy, popular input, and rights-based frameworks in favor of top-down-control—heightens the urgency of this task. 

The following analysis, drawn from contributions at the Forum’s May 2023 workshop “Closing Knowledge Gaps, Fostering Participatory AI Governance: The Role of Civil Society” in Buenos Aires, Argentina, presents some initial reflections from expert stakeholders—chiefly within the digital rights and open government communities—about the present state of the AI governance landscape and potential avenues for civil society intervention.

Several key points emerged from our discussion including. First, stakeholders across government, media, civil society, and the private sector need to see the human agency, relationships, and structures behind AI models—whether the social inequalities that produce biases in data and design, or the political relationships that underpin surveillance deals. Recognizing the human factors and choices that determine how AI systems affect us, rather than seeing these impacts as inevitable, is critical to maintaining democratic accountability for the policy makers, developers, and others who exercise power over and through AI technologies.

Second, these contributions underscore the urgency of equipping democratic societies and institutions to keep up with a constantly evolving set of AI harms and risks. The absence of established norms, learning processes, and institutions addressing AI harms and risks creates challenges for both governments seeking to regulate AI and civil society organizations considering how to use the technology responsibly. Even as processes such as algorithmic impact assessments (AIAs) become more institutionalized, the underlying technical landscape is changing.

Finally, participants stressed the importance of developing new strategies, processes, and collaborations to give real force to principles such as AI transparency, accountability, and privacy by design. Participants faced serious challenges both in engaging with the private-sector actors responsible for much of the decision-making around AI, and in translating state transparency and accountability mechanisms into meaningful rights protections. Deeper involvement by affected stakeholders, especially marginalized communities, at earlier stages of regulatory and design processes was a recurring demand.

As AI use grows more pervasive, our expectations of privacy, access to public goods, and opportunities to challenge injustice from the courtroom to the workplace are likely to increasingly depend on the rules and norms we establish for AI systems. If we are to ensure that these choices reflect the full range of social, civic, and human rights concerns at stake, civil society will have an important role to play in determining how democratic societies use and live with AI.