Publications

NIHRC response to the NI Executive Draft AI Strategy

Key messages 

Human rights by design: AI should be developed, deployed, and monitored with human rights safeguards built in from the outset. AI systems should not be used until those safeguards are in place and should be withdrawn where unacceptable risks emerge.

Meaningful human oversight: Human decision makers must remain accountable and exercise independent judgement over AI outputs. Public authorities remain responsible for complying with the European Convention on Human Rights (ECHR) even when using AI under Section 6 of the Human Rights Act 1998.

Accountability and redress: The strategy should provide for effective, accessible redress mechanisms for people harmed by AI decisions. Meaningful transparency is not simply a technical specification, it is essential for accountability, procedural fairness and the effective protection of human rights.

Data protection: Protection of personal data is a fundamental right linked to Article 8 ECHR. Applicable EU data protection standards remain relevant in Northern Ireland through Windsor Framework Article 2.

Fairness and transparency: The NIHRC highlights the usefulness of the Council of Europe Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law when developing the AI Strategy and calls for explicit human rights compliance measures throughout AI development and deployment. Human rights compliance must be central to all AI use, and the AI Strategy should explicitly embed this principle.

Environmental sustainability: The NIHRC appreciate the recognition in the AI Strategy of AI's environmental impacts, including energy consumption, water use, emissions, and data centres. The right to a safe, clean, healthy and sustainable environment is an internationally recognised human right. The NIHRC recommends democratic scrutiny and community engagement on major AI infrastructure projects.

Societal benefit: The strategy should require human rights impact assessments and due diligence, not just cost-benefit analyses. The NIHRC highlights the Council of Europe's HUDERIA methodology as a useful model.

Training and accessibility: The AI Strategy should address risks such as unemployment, precarious work, and digital exclusion. Accessibility for persons with disabilities must remain a central consideration.

Public procurement: AI procurement processes should explicitly require evidence of human rights compliance. Procurement officials should be given the expertise needed to assess AI systems effectively.

Participation and engagement: Human rights and equality bodies should be included among stakeholders consulted on AI governance, alongside civil society and other groups.

Business and human rights: Private companies developing, procuring, or deploying AI should conduct human rights due diligence in line with the UN Guiding Principles on Business and Human Rights.