Source-linked AI summary
Designing for Human Rights in AI
Evgeni Aizenberg, Jeroen van den Hoven
TL;DR
AI systems increasingly affect human rights, while existing technical approaches often lack empirical attention to societal context and stakeholder input. The paper uses Design for Values, Value Sensitive Design, and Participatory Design to propose a stakeholder-engagement roadmap for translating human rights into context-dependent design requirements. It argues that this approach can support individuals’ dignity and engineers’ acceptance and trust, while acknowledging normative and regional limitations.
Problem
AI systems create socio-ethical challenges affecting fundamental human rights, but technical approaches often lack societal-context analysis and stakeholder input.
Method
The paper presents a Design for Values roadmap drawing on Value Sensitive Design and Participatory Design to engage stakeholders in defining context-dependent requirements.
Results
The approach is presented as supporting individuals whose dignity is paramount and engineers whose products enjoy greater acceptance and trust.
Takeaways & Limitations
Designing for human rights through proactive stakeholder engagement is presented as key to achieving a changing socio-technical landscape in which dignity remains paramount.
Takeaways & Limitations
The analysis recognizes inherent normative biases and calls for follow-up studies at a more regional level.
Abstract
from arXiv · showhide
In the age of big data, companies and governments are increasingly using algorithms to inform hiring decisions, employee management, policing, credit scoring, insurance pricing, and many more aspects of our lives. AI systems can help us make evidence-driven, efficient decisions, but can also confront us with unjustified, discriminatory decisions wrongly assumed to be accurate because they are made automatically and quantitatively. It is becoming evident that these technological developments are consequential to people's fundamental human rights. Despite increasing attention to these urgent challenges in recent years, technical solutions to these complex socio-ethical problems are often developed without empirical study of societal context and the critical input of societal stakeholders who are impacted by the technology. On the other hand, calls for more ethically- and socially-aware AI often fail to provide answers for how to proceed beyond stressing the importance of transparency, explainability, and fairness. Bridging these socio-technical gaps and the deep divide between abstract value language and design requirements is essential to facilitate nuanced, context-dependent design choices that will support moral and social values. In this paper, we bridge this divide through the framework of Design for Values, drawing on methodologies of Value Sensitive Design and Participatory Design to present a roadmap for proactively engaging societal stakeholders to translate fundamental human rights into context-dependent design requirements through a structured, inclusive, and transparent process.
1. INTRODUCTION
AI systems increasingly affect fundamental human rights, yet technical responses to their socio-ethical challenges often omit societal context and stakeholder input. The paper proposes a Design for Values roadmap to translate human rights into context-dependent requirements through inclusive stakeholder engagement.
- Algorithmic systems increasingly shape decisions across domains including employment, policing, credit, insurance, and healthcare, creating major socio-ethical challenges.
- These technological developments are consequential to fundamental human rights and values including human dignity, freedom, and equality.
- Engineering responses to complex socio-ethical problems often lack nuanced empirical study of societal context, stakeholder needs, and affected stakeholders’ values.
- Ethically and socially aware AI proposals often stress transparency, explainability, and fairness without specifying how to proceed beyond those principles.
- The paper shifts from a solutions-oriented to a process-oriented approach that includes societies, social actors, institutions, and interactions.
- Drawing on Design for Values, Value Sensitive Design, and Participatory Design, the authors present a structured, inclusive, and transparent roadmap for translating human rights into context-dependent design requirements.
2. DESIGN FOR VALUES
Design for Values translates abstract moral and social values into context-dependent norms and socio-technical design requirements through stakeholder-informed reasoning. The paper combines iterative conceptual, empirical, and technical investigations to support transparent debate and socially aware design.
- Design for Values explicitly translates moral and social values into context-dependent design requirements, including both social and technical requirements.
- Value specification expands values into norms and then into implementable socio-technical requirements, as privacy can involve consent, confidentiality, and erasure.
- The values hierarchy is context-dependent rather than deductive, allowing stakeholders to debate whether proposed interpretations are sufficient or relevant.
- Stakeholder-formulated reasoning connects lower-level requirements to higher-level values and supports inclusive, cross-disciplinary discussion of socio-ethical implications.
- A structured representation of moral reasons can support technical improvement, social legitimacy, and trust building between parties.
- Value Sensitive Design integrates conceptual investigations of stakeholders and values, empirical investigations of stakeholder experiences, and technical implementation and evaluation.
- These investigations are iterative and integrative, while surveys, interviews, observation, card methods, prototyping, and value scenarios elicit stakeholder concerns and longer-term implications.
3. HUMAN RIGHTS AS TOP-LEVEL REQUIREMENTS
The roadmap adopts human dignity, freedom, equality, and solidarity as top-level requirements, then examines their implications across AI contexts and stakeholder concerns. It places dignity at the root and relates other values to protecting it from AI-related harms.
- The roadmap adopts human dignity, freedom, equality, and solidarity from the EU Charter as top-level requirements guiding AI design.
- The paper examines how Charter values apply to autonomy over self-representation, non-discrimination, privacy, transparency, and automation’s job-market effects.
- 3.1 Dignity: Human dignity is treated as the foundational, overarching value and placed at the root of the design roadmap.
- 3.1 Dignity: The roadmap connects dignity to freedom, equality, solidarity, and the right to life, with the latter linked directly to safety in AI contexts.
- 3.2 Freedom: AI systems can undermine autonomy when statistical inferences judge people by predicted traits rather than their actions and intentions.
- 3.2 Freedom: Algorithmic targets such as criminal risk, personality, job performance, and credit risk are fluid and contestable, but systems reduce them to quantifiable inputs and outputs.
4. FROM ROADMAP TO PRACTICE, MOVING FORWARD
The roadmap is intended to move from human-rights principles toward practical, context-specific design through iterative stakeholder engagement. Its implementation raises unresolved questions about stakeholder selection, consensus, certification, and institutional responsibility.
- The roadmap leaves practical questions about selecting stakeholders, assessing consensus, certifying integrity, verifying iterations, and adjudicating unresolved disputes.
- Implementation requires case studies that test stakeholder-engagement methods, context-dependent requirements, socio-technical solutions, and human-AI interactions.
- The authors envision progressing from case studies toward design protocols and certification procedures applicable across social domains.
- Design for Values complements human-rights due diligence by translating stakeholders’ positive needs and values into concrete, inclusive design requirements.
- Values hierarchy mapping links system components to normative requirements, making stakeholder discussions across abstraction levels more tractable and transparent.
- Companies can use the approach to study societal context, understand stakeholder needs and human-rights implications, and formulate specifications with affected stakeholders.
5. CONCLUSION
The paper proposes Designing for Human Rights in AI: translating human-rights values into structured, inclusive, and transparent socio-technical design requirements. It emphasizes regional context, stakeholder participation, and avoiding the assumption that complex societal problems can be solved by technology alone.
- AI systems raise major socio-ethical challenges, including discrimination, unjustified action, and job-market effects of automation.
- The paper translates human-dignity, freedom, equality, and solidarity values into normative requirements for a structured, inclusive, and transparent design process involving diverse stakeholders.
- Designing for Human Rights treats regional context as consequential, focusing on the EU while acknowledging that human-rights interpretations vary across regions and that the analysis has normative biases.
- The authors reject imposing the EU view on other regions and present the work as an example of an inclusive process that could be applied in other regional contexts.
- A joint socio-technical perspective is needed because technology may advance human rights, but complex societal problems cannot be fixed by smart technology alone.
- The approach aims to support inclusive debate about AI’s role in advancing human rights and to develop human–AI interactions consistent with moral and social values.
- The paper identifies communication across disciplines, stakeholder value tensions, and recognizing that more automation is not always desirable as practical challenges.
- The anticipated benefits extend to individuals whose dignity is prioritized and engineers whose products may gain greater acceptance and trust.