Pakistan’s Draft National AI Policy: fostering responsible adoption and economic transformation

Sahar Iqbal

Akhund Forbes, Karachi

sahar.iqbal@akhundforbes.com

Abstract

In May 2023, Pakistan’s Ministry of IT & Telecom introduced a Draft National AI Policy ('the Policy') as part of its Digital Pakistan vision.[1] This Policy aims to transform Pakistan into a knowledge-based economy and create a conducive ecosystem for the responsible adoption of AI. Aligning with global initiatives, the Policy emphasises the ethical and responsible use of AI, encourages investment in research and development infrastructure, addresses job displacement concerns, and essentially discusses leveraging AI for economic growth. This article explores the Draft National AI Policy, highlighting its objectives, targets, implementation framework, and the significance of addressing ethical considerations.

The Biden administration propagates AI regulations

Washington has been at the forefront of discussions surrounding AI regulations, recognising the need for accountability measures to address the impact of AI on national security, education, and society at large. The Biden administration has sought public input on potential measures to ensure the ethical and responsible use of AI systems, emphasising the importance of safety and addressing concerns relating to AI’s inherent political biases, deception and privacy. As the discussions continue and regulatory measures are being considered, Pakistan has also devised a Draft National AI Policy, aligning with global initiatives, and focusing on responsible AI adoption.

Pakistan’s vision for digital transformation

Pakistan’s Ministry of IT & Telecom is dedicated to realising the 'Digital Pakistan' vision,[2] which aims to provide timely and equal access to opportunities through propagating a culture of innovation. The introduction of the Draft National AI Policy marks a pivotal milestone in this process, with the goal of transforming Pakistan into a knowledge-based economy. The Policy seeks to create an enabling environment for the responsible use of AI, stimulate growth, identify new areas of development, and address the evolving job market.

Objectives and targets of the Draft National AI Policy

Significantly, the Policy recognises the disruptive nature of AI and its potential for socio-economic restructuring. To ensure a focused approach, the Policy sets out specific targets to be achieved by 2028. These targets include widespread public awareness of AI and allied technologies, developing skilled human capital, upskilling the existing workforce, integrating AI in education, establishing a National AI Fund, creating Centres of Excellence in AI and Allied Technologies, transforming the public sector through AI, etc. These targets aim to propel Pakistan’s AI development and recognition on the global stage, while ensuring responsible and inclusive AI adoption.

Regarding socio-economic restructuring, the Policy significantly provides for the empowerment of women and people with disabilities (PWDs) through carrying out AI-training programmes. As part of the National High-Impact Skill Development Programme, a branch that focuses on engaging marginalised women and PWDs through specialised coursework and educational methods is to be established, which is provided for in the Policy. The implementation of this programme will involve collaboration with institutions such as the Bait-ul-Maal, Ehsaas Programme, and other organisations which have access to these marginalised groups. The programme will also be promoted through the DigiSkills.pk platform to ensure maximum inclusivity across the country. Additionally, a special quota will be reserved for marginalised women and PWDs in each cohort.

Funding for implementing the policy

Furthermore, the Policy provides for the establishment of the National Artificial Intelligence Fund (NAIF). This is an initiative of the Ministry of IT & Telecom in Pakistan, aimed at promoting research and development in AI. The establishment of the fund will be mandated by the Ministry, and will receive a portion (not less than 30 per cent) of funds from the Research & Development Fund on a perpetual basis. The Ministry is to allocate an initial working capital through PSDP funds to support the initiative in the first two years. The NAIF will operate as an autonomous organisation with an independent board of directors, comprising industry and academia experts, representatives from Ignite R&D Fund, and government officials. The fund will have the authority to raise additional capital through international grants and partnerships, facilitate collaborations with local and international organisations, and incubate R&D initiatives and startups. The implementation and management of NAIF will be overseen by a chief executive officer recruited from the industry, ensuring transparency, accountability and adherence to the Policy’s directions.

Implementation framework and review procedure

The Policy establishes a comprehensive framework which includes a steering/management committee, working groups, a policy implementation cell, and a review procedure with specific timelines.

The steering/management committee, consisting of representatives from government, academia, industry and civil society, will oversee and guide the implementation of the policy. It will provide guidelines for new and existing initiatives and collaborate with international organisations and academia to keep the AI policy up to date.

Under the Policy, working groups representing different policy drivers are to collaborate with international organisations and academia to assist in research areas where policy measures are needed and provide recommendations to the steering committee. Furthermore, to ensure accountability and continuous improvement, the policy includes a review procedure. A comprehensive policy review will be conducted every three years, while the steering committee will review progress and implementation every six months. Working groups will meet quarterly to suggest recommendations, and the policy implementation cell will operate under the Ministry of IT & Telecom.

Addressing ethical challenges

In the US, a major concern being addressed at discussions under the Biden administration is of the ethical issues posed by AI. In alignment, the Policy in Pakistan recognises the ethical challenges associated with AI adoption and emphasises the need for a trustworthy environment. It aligns with global best practices and emphasises the responsible use of AI while safeguarding user data, upholding privacy rights, and addressing concerns related to AI-biases and data-safety.

Addressing issues pertaining to AI’s political biases is of great concern,[3] as researcher David Rozado conducted four political orientation tests on GPT-4 and all results indicated a predominantly progressive inclination. Despite OpenAI’s efforts to enhance the political neutrality of their latest GPT model, Rozado highlights the immense challenge of accomplishing this due to the inherent bias in the training data. Although AI currently available to the public is in its early stages of development, future utilisation, especially in a court of law, will require strict moderation which aligns with a centric approach. Therefore, it is recommended that the draft Policy further addresses issues regarding AI’s political biases and provides for regulatory mechanisms to mitigate such risks.

Conclusion

The Policy represents significant strides by Pakistan towards transforming the country into a knowledge-based economy and fostering responsible AI adoption while outlining specific targets to be achieved within a defined timeline. The implementation framework and review procedure, along with the NAIF, will ensure effective policy execution and continuous improvement. Pakistan’s first-of-its-kind, detailed, and comprehensive draft policy is commendable, as it has the potential to enhance the country’s legal landscape with cutting-edge legislation. By addressing ethical challenges and fostering a trustworthy environment, the Policy aims to harness the potential of AI while safeguarding the interests of citizens.