Pakistan National Artificial Intelligence Policy 2025

What it is

The National Artificial Intelligence Policy 2025 is Pakistan’s first national AI policy approved by the federal cabinet. The Ministry of Information Technology and Telecommunication (MoITT) prepared it after a consultation draft in 2023, and the cabinet approved it on 30 July 2025. The policy text says it fits with the draft Personal Data Protection Act, the Pakistan Cloud First Policy, the Digital Pakistan Policy and the Digital Nation Pakistan Act.

Its vision is an AI ecosystem where AI is used responsibly and ethically to protect individuals and drive inclusive growth while preserving human rights and the rule of law. The policy has development objectives, covering skills, infrastructure, startups and research, and ethical objectives. The ethical objectives commit to fairness, transparency and accountability, to protecting personal data, privacy and security, and to aligning AI with the UN Sustainable Development Goals and UNESCO’s Recommendation on the Ethics of AI.

It is a policy, not legislation. There is no penalty for ignoring it. Its weight for organisations comes from the guidelines, sandboxes and public sector practice it sets in motion.

Who it applies to

Most of the policy commits the Government of Pakistan to act. It sets up a National AI Fund, Centres of Excellence in AI, an Innovation Fund and a Venture Fund, and it tasks national and provincial IT boards with sector roadmaps.

Three groups feel it most directly.

• Federal and provincial public bodies. They are the subject of the commitments on human oversight, the public register of public sector AI systems, impact assessments and audits, and the target that every public servant is trained on AI basics and personal data protection by 2027.
• Companies offering AI tools and services. Section 2.1 says local and international private sector entities offering AI-led tools and services for data controlling and processing will be given instructions and guidelines on data organisation.
• Organisations in regulated sectors. The policy has Centres of Excellence assisting regulators to run sandboxes for the responsible and ethical use of AI, and an open-source AI governance framework that would audit open-source AI use in data-sensitive organisations such as banks.

What it requires

The policy is organised into six pillars.

1. AI Innovation Ecosystem. Funding, Centres of Excellence in Islamabad, Karachi and Lahore with auxiliary centres in other cities, data standards and centralised datasets for training AI models.
2. Awareness and Readiness. A national awareness programme, skills programmes, scholarships, inclusion of women and persons with disabilities, public sector training, and intellectual property reform.
3. Secure AI Ecosystem. Security guidelines across the AI lifecycle, transparency and human oversight, national data security, identity and access management, an open-source AI governance framework, generative AI guidelines and regulatory sandboxes.
4. Transformation and Evolution. Sector roadmaps for priority areas including education, healthcare, governance, agriculture, manufacturing and energy, and a Ranking Management System to help users evaluate AI solutions.
5. AI Infrastructure. A national AI compute grid, national and provincial AI data repositories, AI hubs and use of public cloud.
6. International Partnerships and Collaborations. Agreements, research collaboration and alignment of Pakistan’s AI regulations and standards with international practice.

The governance commitments sit mainly in the third pillar. Section 3.2 makes five commitments on transparency and human oversight.

• human oversight mechanisms for critical AI operations, particularly in high-risk scenarios
• transparency and disclosure of AI use, including publishing public sector AI systems in a public register
• lifecycle evaluations and impact assessments of high-impact AI systems
• regular audits against ethical and security standards, and a legal framework with penalties for non-compliance
• a framework for auditing AI decision-making, for use by third-party auditors, with redress for unwarranted outcomes

Section 3.1 adds a duty to avoid, prevent and eliminate harms and vulnerabilities throughout the AI lifecycle. Section 3.7 tasks a proposed AI Directorate with regulatory guidelines on disinformation, data privacy breaches and fake news from generative AI, and with measures on copyright, such as licensed datasets and attribution. The fourth pillar sets a target to implement an AI maturity model and project management guidelines in at least 50 institutions by 2026.

Implementation runs through an AI Council chaired by the Federal Minister for IT and Telecom, with federal secretaries, provincial chief secretaries and experts from academia, industry and civil society. A Policy Implementation Cell in MoITT manages day-to-day work.

Key dates

• 9 September 2024. The Regulation of Artificial Intelligence Bill 2024, a private member’s bill, is introduced in the Senate and referred to committee. The Senate records that it later stood withdrawn from the committee.
• 30 July 2025. The federal cabinet approves the National Artificial Intelligence Policy 2025.
• 9 February 2026. The Islamabad AI Declaration is presented at the Indus AI Summit, setting national principles that include human accountability under constitutional authority and explainable, auditable and safe AI.
• 26 June 2026. MoITT releases the draft National Data Governance Policy 2026, which would set risk-tiered rules for AI and automated decisions in federal public bodies.

How it relates to other frameworks

The policy’s ethical objectives point to UNESCO’s Recommendation on the Ethics of AI, and its principles sit close to the OECD AI Principles. Like most national AI policies, it says what good looks like and leaves the method to follow-on instruments.

The most concrete of those is the draft National Data Governance Policy, which would require pre-production registration, explainability, drift monitoring, human review on request and a public registry for high-risk automated decision systems in federal public bodies. It reads as the operational version of section 3.2 for government.

The sixth pillar commits Pakistan to align its AI regulations with international standards. For an organisation building its programme now, ISO/IEC 42001 and the NIST AI RMF supply the management system and risk functions the policy assumes, and a current AI system register is the starting point for the public register and the impact assessments it describes.

A note on this page

This is Aicura’s reading of the National AI Policy 2025 as published by MoITT. It is not legal advice and it is not the policy itself. Read the primary source on the MoITT website, and take advice on how Pakistani policy and law apply to your organisation.