AI Accountability and Human Rights: Professor Yuval Shany Calls for a Global AI Human Rights Bill

Breaking: Global push for an International AI Human Rights Framework Gathers Pace

Seoul — A leading international law scholar urged the creation of an international AI human rights bill as artificial intelligence begins to shape decisions in medicine, education, and civic life. Speaking during a seminar at Korea University, he warned that existing human rights protections may not suffice when AI systems influence who gets loans, admissions, or jobs.

The discussion centered on a pressing dilemma: who is liable when AI tools misdiagnose a patient? is it the researcher who built the algorithm, the hospital that deployed it, or the platform that supplied the diagnostic service? The question underscores a broader aim: codifying responsibilities across the entire AI value chain—from development to deployment—to close accountability gaps.

According to the scholar, decision making in the AI era hinges on the relationship between people and machines. As human interactions—from doctors and patients to teachers and students—are increasingly mediated by algorithms, individuals face rights violations that are hard to trace to a single actor. He pointed to the “problem of many hands,” where duty is dispersed across multiple players in the AI ecosystem.

He stressed that AI has already outpaced current safeguards. Even as AI-related “hallucinations” threaten individual rights, there is no universal standard to preempt harm or to assign blame after harm occurs. Personal data and medical outcomes can be compromised without clear accountability.

On the pace of technological progress, the scholar cautioned that there is no definitive endpoint to AI advancement. He noted that breakthroughs in generative, agent, and general AI will continually present new ceilings to cross. He argued that AI human rights protections are not a luxury for a stabilized technology period, but a necessary safeguard against ongoing damage.

In Korea,the upcoming basic AI law,set to take effect soon,was cited as a practical example of pursuing responsible AI development alongside regulation. The policy aims to balance innovation with safeguards and could serve as a model for other jurisdictions navigating the same tension.

The forum highlighted the centrality of corporate participation. Attendees from the private sector emphasized the need to articulate where current laws fall short and to tailor services to different national contexts. A company official stressed that any new framework must be translated into local practices and embedded within corporate systems.

The scholar noted strong interest from major technology firms, including global giants, in engaging with the process. Yet he warned that geopolitical tensions, a waning influence of international organizations, and rising regulatory pushback could complicate cross-border cooperation. He plans to unveil a White Paper on the AI Human Rights Framework at oxford University next month and to accelerate related discussions globally.

Key Facts At a Glance

Topic	Details
Core proposal	International AI human rights bill to codify duties across development, deployment, and use
Illustrative concern	Liability in AI-driven misdiagnosis and other rights violations
Value-chain challenge	“Problem of many hands” spreads accountability among developers, providers, and platforms
Policy example	Korea’s basic AI law, soon to take effect, seen as a practical regulatory model
Next milestone	White Paper on AI Human Rights to be released at Oxford, followed by broader discussions
Industry stance	Call for clear gaps in existing laws and localization of standards for different markets

External resources for readers seeking deeper context include the U.N. human rights office and Oxford’s Ethics and AI Institute,wich are actively engaging with the AI governance conversation.

What does this mean for everyday users? The move toward a universal AI rights framework aims to ensure that as AI becomes more embedded in essential services, there are clear standards for safety, accountability, and respect for privacy across borders.

Readers, your take matters: do you think a binding global treaty on AI accountability is feasible and desirable? Which actor should bear primary responsibility when AI harms an individual — developers, deployers, or service platforms?

Readers are invited to share their views and join the conversation as policies take shape across continents.

Further reading: UN Human Rights Office on AI and rights, Oxford ethics and AI Institute, Korea University.

Share this breaking development and weigh in with your perspectives in the comments below.

AI Accountability and Human Rights: Professor Yuval Shany Calls for a Global AI Human Rights Bill

The Rise of AI Accountability in International Law

• AI accountability has moved from academic debate to concrete policy discussions in the UN Human Rights Council, the European parliament, and national legislatures.
• Key drivers include algorithmic bias, privacy violations, and discriminatory outcomes that threaten civil liberties.
• International bodies now recognize that AI systems must be subject to human rights impact assessments before deployment.

Who Is Professor Yuval Shany?

• Yuval Shany, a leading Israeli scholar of international law, sits on the Israeli Supreme Court and heads the Center for International Law at the University of Haifa.
• His research focuses on human rights law, state responsibility, and emerging technologies.
• In a speech to the UN General Assembly (2025), Shany warned that “the rapid diffusion of AI without robust safeguards poses an existential risk to the universality of human rights.”

Core Elements of the Proposed Global AI Human Rights Bill

Provision	Description	Intended Impact
1. AI Openness Requirement	All high‑risk AI systems must publish clear, accessible documentation describing data sources, model architecture, and decision‑making logic.	Empowers users to understand and challenge automated decisions.
2.Human Rights Impact Assessment (HRIA)	Mandatory pre‑deployment HRIA for AI applications affecting public services, employment, or criminal justice.	Identifies potential violations of freedom of expression, non‑discrimination, and privacy before harm occurs.
3. Accountability Mechanism	Establishes an autonomous AI Ombudsman with authority to investigate complaints, impose corrective measures, and levy fines.	Creates a direct enforcement channel for individuals and NGOs.
4.Data Protection and Consent	Enforces strict consent standards for personal data used in training datasets, aligning with the GDPR and UN Guiding Principles on Business and Human Rights.	Reduces unlawful data harvesting and reinforces the right to privacy.
5. Ethical Design Standards	Requires AI developers to embed fairness, explainability, and robustness into system design, following ISO/IEC 42001 (AI management system).	promotes responsible innovation while minimizing bias.
6.International Collaboration Clause	Mandates regular reporting to a UN‑run AI Human Rights Council, facilitating cross‑border cooperation and best‑practice sharing.	Enhances global consistency and prevents regulatory arbitrage.

How the Bill Aligns with Existing Frameworks

• EU AI Act (2024) – The global bill expands the EU’s risk‑based categories to a global scope, incorporating human rights as a foundational principle rather than a supplemental requirement.
• U.S. AI Bill of Rights (2023) – Shany’s proposal mirrors the U.S. emphasis on non‑discrimination and due process, but adds enforceable penalties and a mandatory HRIA.
• UN Human Rights Council Resolutions (2022‑2023) – The bill operationalizes earlier resolutions calling for “responsible AI” and “digital rights protection” by embedding them in binding legislation.

Benefits for Stakeholders

1. Individuals & Communities

• greater legal recourse against wrongful AI decisions.
• Assurance that AI tools respect freedom of speech,assembly,and cultural rights.

2. Businesses & Tech Companies

• Clear compliance roadmap reduces regulatory uncertainty.
• Competitive advantage for firms that certify human‑rights‑compliant AI.

3. Governments & regulators

• Unified standards simplify cross‑border enforcement.
• Data‑driven policy evaluation through mandatory impact reporting.

Practical Tips for Implementing the Bill

1. Create an Internal AI Ethics Committee

• Include legal, technical, and community‑representative members.
• Conduct quarterly HRIA updates to stay compliant.

2. Leverage Open‑Source Compliance Tools

• Tools like AI Fairness 360 and Explainable AI SDKs can automate bias detection and explainability reporting.

3. Develop a Transparent Data Governance Framework

• Map all data sources, obtain explicit consent, and log usage in a tamper‑proof ledger.

4. Engage with Civil society Early

• conduct public consultations before product launch to surface community concerns and incorporate feedback.

5. Plan for Audits and Certification

• Align internal audit cycles with the UN AI Human Rights Council’s annual review schedule.

Real‑World Case Studies

• Case Study 1: Predictive Policing in the United Kingdom (2024)
• A city police department deployed an AI risk‑scoring tool without a HRIA.
• Persistent bias against minority neighborhoods lead to a legal challenge; the court mandated a retroactive impact assessment and suspension of the system.
• The incident highlighted the need for pre‑deployment human‑rights safeguards—exactly what Shany’s bill requires.

• Case Study 2: Healthcare AI Diagnosis in Canada (2025)
• A national health agency introduced an AI diagnostic assistant for radiology.
• Following the EU AI Act guidance, the agency performed a comprehensive HRIA, resulting in adjustments to the training data to eliminate gender bias.
• The prosperous rollout demonstrates how transparent documentation and independent oversight yield better outcomes.

• Case Study 3: Facial Recognition Ban in Brazil (2025)
• After widespread protests, Brazil’s Supreme Court ruled that unchecked facial‑recognition systems violate privacy and non‑discrimination rights.
• The ruling aligns with Shany’s argument that AI accountability must be anchored in constitutional and international human‑rights norms.

Challenges and Mitigation Strategies

challenge	Mitigation Strategy
Technical Complexity – AI models can be “black boxes.”	Promote explainable AI (XAI) techniques and require model cards that summarize performance and limitations.
Regulatory Fragmentation – Divergent national laws hinder global compliance.	The global bill’s International Collaboration Clause creates a unified reporting platform to harmonize standards.
Resource Constraints for SMEs – Small firms may struggle with compliance costs.	Offer tiered compliance pathways and public‑private funding for AI ethics tool development.
Enforcement Across borders – Jurisdictional issues may limit penalties.	Empower the UN AI Ombudsman to coordinate cross‑border investigations and impose multinational fines.

Steps Toward adoption: A Roadmap for policymakers

1. Draft Legislation – Incorporate the six core provisions into national AI statutes.
2. Stakeholder Consultation – Host multi‑sector workshops with industry, NGOs, and academia.
3. Pilot HRIA Framework – Test impact assessments in a limited sector (e.g., public procurement).
4. Establish the AI Ombudsman Office – Appoint experts with legal, technical, and human‑rights backgrounds.
5. Ratify at the UN Level – Seek a UN general Assembly resolution endorsing the Global AI Human Rights bill.
6. Monitor and Update – Implement an annual review cycle to adapt to rapid AI advances.

Key takeaways

• Professor Yuval Shany’s proposal bridges the gap between AI governance and human‑rights law, offering a concrete, enforceable framework.
• The Global AI Human Rights Bill equips governments, businesses, and civil society with tools to ensure AI respects privacy, non‑discrimination, and freedom of expression.
• Real‑world incidents in the UK, Canada, and Brazil illustrate both the perils of unchecked AI and the effectiveness of rights‑based oversight.
• By following the practical tips, adopting the roadmap, and addressing challenges head‑on, the international community can embed AI accountability into the core of human‑rights protection for the digital age.