It is no hidden fact that artificial intelligence (AI) has already impacted the industry, human capital, and society at large in many ways. AI technologies and their applications will continue to proliferate in terms of their evolution and usage across human life.

The global market size of AI-related products is expected to grow from $454 billion in 2022 to $2.58 trillion by 2032. The technology is lucrative enough to be considered a harbinger of change and economic value creation. Hence, countries around the globe are pooling their wisdom and putting their resources to reap the benefits of a multitude of avenues.

Pakistan has a draft AI policy in 2022, a rather delayed but marvellous attempt to steer the country towards the path to progress while adopting AI. The policy envisions attaining a hybrid intelligence ecosystem for the equitable, responsible, and transparent use of AI.

Policy targets are devised from two agenda streams: developmental and responsibility. However, they are silent on the adoption of AI in governance and public service delivery.

Target 13 implicitly describes the access to data sets and sandboxes under the head of the transformation of the public sector through AI and allied technologies. However, vested details are missing. The AI enablement pillar of the policy mentions various initiatives like establishing a centre of excellence of AI (CoE-AI), an AI fund and catalysing social development through AI.

However, the question remains have we learned about the use of information and communications technology research and development funds and the way National Technology Fund Ignite has used such funds?

The proposed funding mechanism relies on the public sector development programme, which is prone to funding cuts at any time. Therefore, the funding mechanism proposed is not promising. Similarly, perpetual funding progression methods are not eloquent.

As far as the establishment of CoE-AI is concerned, we already have four such centres already, one for AI, quantum, data, and cyber security. To date, we have not been able to harness their research and innovation capability. How about building on the capabilities which can be offered by these centres and avoiding re-invention of the wheel (as proposed in the draft policy)?

Transformation of the social sector is a huge need of our country, and AI has all the potential to help us navigate the crisis of education, health, traffic congestions, city management, agriculture, supply chain, and the list is long. Furthermore, AI has a lot to offer to improve our crippling governance.

The AI policy draft 2022 sometimes gets deeper into the fields of health and education and, at times, touches upon agriculture and the value chain superficially. We need to arrive at our specific needs at the federal and provincial levels across various public utilities and then we can carve out a policy priority around the use of AI in these domains.

We need a specialised workforce for preparing AI solutions and adopting available AI platforms and development flavours in relation to our needs in social and economic sectors. The reality is that we don’t have the required capacity in academia to train the workforce needed, and our industry is not ready to absorb it.

Technical knowledge and competence required by academia to train an able workforce is another missing link. There is a decisive differential in pay and incentives if the potential workforce opts to work abroad or online versus getting employed by local companies.

How does the government want to address this challenge? The policy options narrated in the draft are not clear, and the required viability of proposed options is yet to be worked out.

The initiatives of up-skilling, re-skilling, and academic scholarships are proposed in the draft as well. However, the numbers proposed in the policy for academic scholarships and the nature of scholarship arrangements need to be revisited.

Data and computational power are the backbones of any AI utility. It remains a fact (and rightly mentioned in the AI draft policy 2022) that our data is highly fragmented. On the other hand, we have not been able to achieve the basics of data availability, security, sharing, and pooling.

In the absence of big data governance, protection, and implementation of protocols, it takes a lot of work to think of introducing and using AI.

Similar is the case with computational power. We lack the fundamentals for servicing and timely repair of personal computers, and our government is far away from the infrastructure of e-governance. In this situation, how are we planning to build computational agility, and how do we want to acquire and use this power for AI? The draft policy has not been able to address these areas adequately.

Similarly, we can’t protect our people from cyber-attacks that compromise their bank accounts every now and then because our existing infrastructure is working in silos. What will enable us to use the data for various AI applications? We need to find the right policy choice for this facet as well.

Defence and security is a domain ignored by the draft AI policy 2022. With the rise of automated and unmanned warfare, this is an area that requires further deliberations and a strategic focus.

The policy implementation and review procedure are quite commendable. In summation, we have taken the right start while introducing the draft AI policy 2022. Certainly, there is an opportunity to make it more viable for reaping economy and society-wide benefits.