Trustworthy AI and Democracy: A Dual Taxonomy and Operational Framework of Democratic Risks and Contributions
Organizations: Dept. of Computer Science and Artificial Intelligence, DaSCI, University of Granada, Spain · Dept. of Philosophy, University of Vienna, Vienna, Austria · School of Engineering, Cornell University, Ithaca, NY, United States · Dept. of Mathematics, Khalifa University of Science and Technology, Abu Dhabi, UAE ·
Abstract
Artificial Intelligence (AI) can undermine and strengthen democratic governance. This paper introduces the AIRD/AIPD framework, a dual taxonomy and operational framework for analyzing AI's democratic risks and conditional contributions. The AI Risks to Democracy (AIRD) taxonomy identifies threats to autonomy, fairness, public discourse, democratic institutions, and trust, while the AI's Positive Contributions to Democracy (AIPD) taxonomy captures AI's potential to enhance efficiency, accessibility, security, informed citizenry, transparency and accountability, inclusive democratic dialogue, and evidence-based policymaking. Grounded in five democratic principles and the seven Trustworthy AI requirements proposed by the European Commission's High-Level Expert Group on AI, the framework connects democratic goals with governance and design mechanisms. It provides an operational process for identifying affected contexts and actors, classifying primary and secondary AIRD/AIPD effects, mapping them to relevant Trustworthy AI requirements, and defining proportionate governance responses and responsible actors. An operational mapping and two worked examples illustrate its application to electoral misinformation and AI-assisted public service allocation. The analysis highlights the transversal relevance of transparency, accountability, fairness, and societal wellbeing, while recognizing trade-offs, implementation costs, differential impacts, institutional capacity, and contextual variation across democratic settings. The framework is conceptual and has not been empirically validated, but it offers a traceable and adaptable basis for research, policy analysis, institutional oversight, and responsible AI design. By connecting democratic diagnosis with governance action, it bridges the gap between normative principles and the operational requirements of trustworthy and democratically legitimate AI.
Figures & tables
| Category | Diagnostic focus | Trustworthy AI requirements | Governance response | Responsible actors |
| AIRD1. Subversion of Autonomy | Manipulation of citizens’ choices | Human agency; technical robustness; transparency | Disclosure, contestability, and anti-manipulation safeguards | Regulators; platforms; providers; auditors |
| AIRD2. Unfairness | Biased outcomes or exclusion from democratic participation | Technical robustness; fairness; societal wellbeing | Bias assessment, representative data, accessibility, and appeal procedures | AI providers; public agencies; auditors; civil society |
| AIRD3. Power Asymmetries | Concentrated informational, economic, or institutional power | Data governance; transparency; fairness; accountability | Audit trails, transparency portals, and independent oversight | Regulators; platforms; AI providers; auditors; civil society |
| AIRD4. Danger of Authoritarianism and Totalitarianism | Use of AI for surveillance, coercion, or centralized political control | Data governance; transparency; fairness; societal wellbeing; accountability | Legal limits on surveillance, human rights impact assessments, independent supervision, and redress mechanisms | Regulators; courts; public institutions; oversight bodies; civil society |
| AIRD5. Undermining Public Discourse | Misinformation, polarization, or degradation of democratic deliberation | Transparency; fairness; societal wellbeing; accountability | Provenance, platform audits, fact-checking, and rapid reporting | Platforms; regulators; civil society; electoral authorities |
| AIRD6. Undermining Democratic Principles | Weakened participation, pluralism, representation, or accountability | Transparency; societal wellbeing | Participatory governance, democratic impact assessments, public consultation, and institutional transparency duties | Public institutions; regulators; electoral authorities; civil society |
| Step | Application |
|---|---|
| Observed Issue | Synthetic media falsely attribute statements to a candidate before an election. |
| AIRD Classification | AIRD5; AIRD7; and, with targeted profiling, AIRD1. |
| Trustworthy AI Requirements | Rq4 Transparency ; Rq7 Accountability ; and Rq3 Privacy and Data Governance . |
| Governance Responses | Provenance labels, ad repositories, rapid audits, reporting channels, and electoral authority–platform cooperation. |
| Step | Application |
|---|---|
| Observed Issue | AI-assisted allocation may improve efficiency but create bias, opacity, weak contestability, and unequal group effects. |
| AIPD Classification | AIPD1 Efficiency ; AIPD7 Evidence-Based Policymaking ; and, where AI reduces access barriers, AIPD2 Accessibility . |
| Associated AIRD Risks | AIRD2 Unfairness ; AIRD3 Power Asymmetries . |
| Trustworthy AI Requirements | Rq1; Rq2; Rq3; and Rq4. |
| Governance Responses | Human oversight, bias and robustness tests, data safeguards, explanations, appeals, outcome monitoring, and independent audits. |
| Req. | Mapped AIRD risks | Freq. |
|---|---|---|
| Rq4 | AIRD1, AIRD3, AIRD4, AIRD5, AIRD6, AIRD7 | 6/7 |
| Rq6 | AIRD2, AIRD4, AIRD5, AIRD6, AIRD7 | 5/7 |
| Rq7 | AIRD3, AIRD4, AIRD5, AIRD7 | 4/7 |
| Rq5 | AIRD2, AIRD3, AIRD4, AIRD5 | 4/7 |
| Rq3 | AIRD3, AIRD4 | 2/7 |
| Rq2 | AIRD1, AIRD2 | 2/7 |
AI Governance
Artificial intelligence governance studies how laws, institutions, and technical controls shape AI development, deployment, accountability, and social effects.
See all 386 papers