Reaping the skills premium

As the AI Spring unfolds, the world of work is turning a new leaf. The air is thick with both excitement and uncertainty – over automation, potential job losses, the future of human creativity and the shifting meaning of work itself. Some brace for pruning, others prepare for growth spurts. One thing is clear: skills are the new currency. In the online freelance market, the skills premium is already blooming.

GenAI’s (Generative Artificial Intelligence) impact is comparable to some of the greatest technological breakthroughs ever. Unlike traditional AI, which focussed largely on process optimisation, GenAI extends into creative and cognitive domains once thought intrinsically human. Beyond improving recommendation algorithms, it can draft emails, design ad campaigns, outline curricula and develop prototypes – turn ideas into tangible output.

Powered by massive datasets and energy-intensive hardware, GenAI offers time and cost-efficient solutions for everyone. It transforms the services and creative industries while reshaping work in the broader context. Naturally, its adoption rate has outpaced any other innovation, driving an eightfold investment increase since Chatgpt‘s launch in November 2022. According to Stanford’s AI Index Report 2025, 71 per cent of businesses globally leveraged GenAI for at least one task in 2024, with developing markets not so far behind at 67 per cent.

The promise of economic gains and leapfrogging potential for resource-constrained players is driving this uptake. McKinsey & Company estimates that GenAI could add over $2.6 trillion to the global economy by 2040. But what does this mean for the future of work? What role will people play, how will they do it, and how will society value it?

The world of work has always evolved in seasons, each brought about by the winds of technological change. Skill specialisation began as early as the Bronze Age, while the Industrial Revolution established the division of labour as the default.

Each shift displaced some jobs (the replacement effect) while creating some new ones (the reinforcement effect), redefining skill demands. The printing press rendered scribes obsolete but created new roles like press operators. Similarly, UNIVAC I’s launch reduced demand for human ‘computers’, many of whom transitioned to emerging fields like FORTRAN programming.

Today, GenAI presents another pivotal moment in the story of work – the next phase of human-machine collaboration. What sets it apart is not just its novelty, but also the nature and scale of disruption it brings.

The World Economic Forum estimates that Artificial Intelligence, driven mainly by GenAI innovation, will displace 92 million jobs by 2030 while creating around 170 million new jobs globally. The scale of this disruption will vary across sectors, roles, and the skills spectrum, though.

An early blueprint of the GenAI-powered work environment is already visible in the online freelance markets. It warrants attention for two reasons.

First, freelance markets are early indicators of emerging industry trends, evolving work roles, and changing skill requirements. Second, skills, not schooling credentials or geography, are the strongest currency in these markets. Their impact extends to emerging markets like Pakistan since they thrive on the globalisation of opportunities and resources.

Initial evidence from the public rollouts of LLMs showed a decline in job postings for tasks easily automatable through structured data analysis, such as writing, basic programming, and web development. Some studies estimated a drop of over 20 per cent in postings for these roles.

Yet it’s not all doom and gloom. Upwork, a leading freelancing platform, reported a 2.4 per cent rise in total postings and a 1.3 per cent increase in per-contract earnings by mid-2023. It highlighted job creation aligned with the emergence of new skill categories like LLM tuning and workflow automation. In fact, freelancers engaged in AI-specific gigs were reportedly earning 40 per cent higher than those doing non-AI work, suggesting that the reinforcement effect is already at play.

The most significant gains are visible for gigs related to technological and digital solutions, followed by business operations. In contrast, creative and service-oriented roles continue to face the replacement effect with fewer jobs and lower pay for traditional copywriters, translators, sales assistants, and other such roles.

That said, a silver lining is emerging. Recent evidence suggests that freelancers upskilling with GenAI-specific or complementary skills, such as AI content creation, fact-checking/evaluation, and responsible AI use, are managing to retain ground in their established fields.

The replacement effect is not absolute. Even in seemingly automatable roles, GenAI’s efficiency declines as task complexity increases. It struggles with tasks that require nuance, analytical depth and contextual specificity. This is where critical thinking, creative problem-solving, socio-emotional intelligence and human experience continue to hold value.

The crux? GenAI is skills-biased and has paradoxical effects. It automates traditionally low-skilled, low-value work while increasing returns for high-skilled, high-value roles. In doing so, it risks widening inequalities across the skills spectrum.

While GenAI promises skill democratistion by real-time learning opportunities, it also requires strong foundational knowledge and domain expertise to tackle complex problems. Once again, its potential and limitations combine to reinforce the growing skills premium.

How much of it can Pakistan reap? That will depend on the groundwork it lays today.

From a bird's-eye view, Pakistan is the fourth-largest freelance market by active gig workers. However, it is not ranked among the top five most valuable talent pools globally. That’s precisely why GenAI’s impact has direct and urgent implications for young Pakistani freelancers who are otherwise facing limited opportunities in the traditional home market.

Currently, only about 1.0 per cent of Pakistani freelancers offer high-value services like artificial intelligence and blockchain. About 55 per cent of the country’s registered active freelance workforce work in web development, graphic design, and content writing. Gigs in these fields have traditionally been low-paying and now have high exposure rates to GenAI-enabled automation.

Beyond the registered platform economy lies a vast, often overlooked informal sector in the Pakistani freelance market. This space has historically provided opportunities to for small towners, women, and students, mainly though outsourcing of low-value but foreign currency-priced gigs, generally for minimal local currency compensation. Weak-tie social networks and informal, unwritten agreements typically mediate these transactions, helping Pakistan maintain a cost advantage in the global freelance markets.

GenAI is, however, shifting the equation in demand-side markets like North America. Therefore, the focus must pivot to building a new advantage: human capital.

To reap the skills premium, investment in skills comes first. This puts the spotlight on governments and educational institutions. Public-private programmes like DigiSkill have made significant contributions to national skill development across domains. However, like other formal initiatives, their adaptability is constrained by institutional interties.

In contrast, bootcamps and micro-credential programmes, offered by edtech platforms, provide more efficient options for re- and up-skilling both freelancers and the traditional workforce. The key now is to realign efforts towards high-value, (Gen)AI-specialised and complementary skills – the very competences defining competitiveness in the new season of work.

Lastly, in the Age of Generative AI, where skills and human adaptability determine who captures the skills premium, individual agency becomes the most crucial element. The secret to sustainable growth lies in learning how to learn, adapt and nurturing the drive to keep evolving.

In a world where even creativity can be automated, the most human skill may well be the ability to regrow – again and again – with every spring.

The writer is a research fellow at the Graduate Institute of Development, Lahore School of Economics.