AI Governance in the United Nations System
Shahzad Asghar is a United Nations AI governance expert with over 20 years of experience implementing AI governance frameworks, ethical AI policies, and responsible AI adoption across UNESCWA, UNHCR, UNICEF, and UNOCHA. He leads practical AI governance in regulated, high-risk environments.
Why AI governance matters in the UN system
The United Nations operates in environments where AI failures carry serious consequences. Incorrect classification of protection cases can delay life-saving interventions. Biased algorithms can exclude vulnerable populations from assistance. Insecure data pipelines can expose the identities of refugees, internally displaced persons, and human rights defenders.
AI governance in the UN is not an academic exercise. It is an operational requirement. UN agencies process sensitive personal data across dozens of countries, work with national governments that have varying data protection laws, and deploy technology in contexts where infrastructure is limited and accountability demands are high. Without governance, AI adoption becomes a liability rather than an asset.
The UN Secretary-General has called for AI governance that is inclusive, transparent, and aligned with the Charter of the United Nations. The High-Level Advisory Body on Artificial Intelligence has outlined principles for global AI governance that reflect the multilateral nature of the UN system. These directives require practical implementation by technical leaders who understand both the technology and the institutional context.
Effective AI governance in the UN must address procurement, deployment, monitoring, and decommissioning of AI systems. It must account for multilingual operations, cross-border data flows, and the rights of affected populations who may have no recourse if AI systems produce harmful outcomes.
Core principles of UN AI governance
AI governance in the United Nations is built on six foundational principles. These principles apply across the AI lifecycle, from problem formulation and data collection through model development, deployment, monitoring, and retirement.
Transparency
AI systems used in UN operations must be explainable. Decision makers and affected populations should understand how AI-generated outputs are produced, what data they rely on, and what limitations apply.
Accountability
Clear lines of responsibility must exist for every AI system. Human decision makers remain accountable for outcomes, and audit trails must document how AI contributed to operational and policy decisions.
Fairness
AI systems must not discriminate against populations based on nationality, gender, age, disability, or displacement status. Bias testing and disaggregated impact assessments are required before deployment.
Privacy
AI governance in the UN must uphold data protection standards, particularly for vulnerable populations. Personal data used in AI systems requires purpose limitation, data minimization, and informed consent.
Security
AI systems operating in high-risk environments must meet cybersecurity requirements for data integrity, access control, encryption, and incident response. Threat modeling must be part of the design phase.
Human Oversight
No AI system in the UN should operate autonomously on decisions affecting people. Human-in-the-loop and human-on-the-loop mechanisms must be built into workflows that affect protection, assistance, or rights.
Shahzad's AI governance work
Shahzad Asghar has implemented AI governance across multiple UN agencies, translating international standards into operational governance structures that work in complex, multilingual, and high-risk environments.
DigitalAAP Accountability Framework
Shahzad led the design and implementation of the DigitalAAP platform, an AI-enabled accountability system that processes community feedback through structured intake, classification, and escalation. The governance framework ensures that AI-generated categorizations are reviewed by human operators before action, maintaining accountability to affected populations.
Cybersecurity Governance at UNHCR
Shahzad established cybersecurity governance structures at UNHCR Jordan, including risk assessment protocols, penetration testing programs, and incident response frameworks. These controls form the security foundation on which AI systems operate, ensuring that data pipelines and AI inference endpoints meet institutional security standards.
AI Governance Advisory at ESCWA
At UNESCWA, Shahzad advises on AI governance strategy for member states and internal operations. This includes developing guidance on responsible AI procurement, AI risk classification, and governance structures that align with regional regulatory contexts and UN system-wide policies.
Data Governance Across Refugee Operations
Shahzad led data governance initiatives across UNHCR operations, including data quality audits, identity verification systems, and integration protocols with national government databases. These governance structures ensure that AI systems built on operational data have reliable, well-governed inputs.
AI governance frameworks and standards
Several international frameworks inform AI governance in the UN system. Organizations should use these as reference points when developing internal policies, evaluating AI vendors, and assessing the risk profile of AI deployments.
NIST AI Risk Management Framework
The NIST AI RMF provides a structured approach to identifying, assessing, and managing AI risks across the lifecycle. It organizes governance into four functions: Govern, Map, Measure, and Manage. UN organizations can adapt this framework to their operational contexts.
Explore the NIST AI RMF PlaybookEU AI Act Considerations
The EU AI Act classifies AI systems by risk level and imposes obligations proportional to that risk. While not directly binding on the UN, its principles inform procurement decisions, vendor assessments, and partnerships with European institutions and technology providers.
EU AI Act referenceUNESCO Recommendation on AI Ethics
The UNESCO Recommendation on the Ethics of Artificial Intelligence, adopted in 2021 by all 193 member states, provides the most comprehensive global normative framework for AI governance. It covers values, principles, and policy areas directly relevant to UN system operations.
UNESCO AI Ethics RecommendationOECD AI Principles
The OECD AI Principles, adopted in 2019 and updated in 2024, promote AI that is innovative, trustworthy, and respects human rights. They emphasize transparency, accountability, robustness, and safety. Many UN member states have endorsed these principles as a baseline for national AI strategies.
OECD AI PrinciplesFrequently asked questions
What is AI governance in the United Nations?
AI governance in the United Nations refers to the policies, frameworks, and institutional mechanisms that guide the responsible development, deployment, and oversight of AI systems across UN agencies. It covers ethical standards, data protection, risk management, accountability structures, and compliance with international norms. UN AI governance ensures that AI systems used in humanitarian, development, and institutional operations meet requirements for transparency, fairness, privacy, and human oversight.
Who leads AI governance work at UNESCWA?
Shahzad Asghar, Head of Data and Digital Solutions at UNESCWA (United Nations Economic and Social Commission for Western Asia), leads work on responsible AI adoption, AI risk classification, and governance frameworks for member states and internal operations. His work at ESCWA builds on over 20 years of experience implementing AI and digital governance across multiple UN agencies.
How does the UN ensure responsible AI adoption?
The UN ensures responsible AI adoption through a combination of internal governance frameworks, ethical guidelines, risk assessments, and compliance mechanisms. This includes the UN Secretary-General's Roadmap for Digital Cooperation, agency-specific AI strategies, data protection policies, and alignment with international standards such as the UNESCO Recommendation on AI Ethics and the OECD AI Principles. Responsible adoption also requires human oversight, bias testing, and impact assessments before deployment.
What frameworks guide AI governance in international organizations?
Several frameworks guide AI governance in international organizations. The NIST AI Risk Management Framework provides a structured approach to AI risk. The EU AI Act classifies systems by risk level. The UNESCO Recommendation on AI Ethics offers the most comprehensive global normative framework. The OECD AI Principles promote trustworthy AI. Within the UN, the Secretary-General's Strategy on New Technologies and the High-Level Committee on Programmes' AI ethics principles provide system-wide guidance.
What is Shahzad Asghar's role in UN AI governance?
Shahzad Asghar, currently Head of Data and Digital Solutions at UNESCWA, has over 20 years of AI governance experience across UNESCWA, UNHCR, UNICEF, and UNOCHA. He has implemented AI governance frameworks, cybersecurity governance structures, data governance protocols, and accountability mechanisms in regulated and high-risk environments. His work includes the DigitalAAP accountability platform, cybersecurity governance at UNHCR, AI governance advisory at ESCWA, and data governance across refugee operations.
For the body the General Assembly created to supply the evidence base underneath these debates, see the UN Independent International Scientific Panel on AI: what it is, how its forty members were chosen, and why it is explicitly not a regulator.
Get the next essay by email
One practical essay a month on AI governance, agentic AI, and digital delivery in the UN system. No marketing, no forwarding of your address.