[Paper Review] A Human Rights-Based Approach to Responsible AI

Vinodkumar Prabhakaran, Margaret Mitchell+2 more|arXiv (Cornell University)|Oct 6, 2022

Ethics and Social Impacts of AI31 citations

TL;DR

The paper argues for grounding responsible AI in a universal human rights framework to align values across cultures, clarify responsibilities, and enable civil society participation. It advocates using UDHR rights to reorient FATE research toward human welfare and rights protection.

ABSTRACT

Research on fairness, accountability, transparency and ethics of AI-based interventions in society has gained much-needed momentum in recent years. However it lacks an explicit alignment with a set of normative values and principles that guide this research and interventions. Rather, an implicit consensus is often assumed to hold for the values we impart into our models - something that is at odds with the pluralistic world we live in. In this paper, we put forth the doctrine of universal human rights as a set of globally salient and cross-culturally recognized set of values that can serve as a grounding framework for explicit value alignment in responsible AI - and discuss its efficacy as a framework for civil society partnership and participation. We argue that a human rights framework orients the research in this space away from the machines and the risks of their biases, and towards humans and the risks to their rights, essentially helping to center the conversation around who is harmed, what harms they face, and how those harms may be mitigated.

Motivation & Objective

• Promote explicit value alignment in AI research using a universal, cross-cultural framework.
• Highlight three core functions of human rights for AI: moral claims, legal regime, and cultural practice.
• Demonstrate how human rights can address alignment, responsibility, and participation gaps in FATE research.
• Bridge the gap between machine learning researchers and civil society through a rights-based vocabulary and participatory processes.

Proposed method

• Articulate a three-fold view of human rights: moral claims, legal regime, and cultural practice.
• Discuss UDHR rights and their relevance to AI harms and governance.
• Illustrate how three specific rights (non-discrimination, health, and share in scientific advancement) translate into design, deployment, and governance considerations for AI.

Experimental results

Research questions

• RQ1 How can a universal human rights framework guide value alignment in AI systems across diverse cultural contexts?
• RQ2 Who bears responsibilities when AI harms occur, and how can rights-based frameworks allocate duties among states, organizations, and individuals?
• RQ3 How can human rights facilitate civil society participation in the design, deployment, and governance of AI technologies?
• RQ4 What concrete implications do UDHR rights have for fairness, health outcomes, and access to scientific benefits in AI applications?

Key findings

• Human rights provide cross-cultural grounding for value alignment in AI, addressing the risks of Western value imposition.
• A rights-based lens clarifies responsibilities across states, organizations, and individuals in AI development and deployment.
• Civil society participation is strengthened when rights holders are foregrounded as participants, not just stakeholders or consumers.
• Rights like non-discrimination, health, and access to scientific advancement yield concrete implications for fairness, healthcare AI, and inclusive innovation.