ARTIFICIAL INTELLIGENCE

“Someone that knows how to use AI will replace someone that doesn’t, even if AI itself won’t replace a person. So getting through the hype to give people the skills they need is critical.” – Andrew Ng – AI guru, Coursera founder

The distinction Andrew Ng draws between AI replacing jobs and AI-capable workers replacing their peers represents a fundamental reorientation in how we should understand technological disruption. Rather than framing artificial intelligence as an existential threat to employment, Ng’s observation-articulated at the World Economic Forum in January 2026-points to a more granular reality: the competitive advantage lies not in the technology itself, but in human mastery of it.

The Context of the Statement

Ng made these remarks during a period of intense speculation about AI’s labour market impact. Throughout 2025 and into early 2026, technology companies announced significant workforce reductions, and public discourse oscillated between utopian and apocalyptic narratives about automation. Yet Ng’s position, grounded in his extensive experience building AI systems and training professionals, cuts through this polarisation with empirical observation.

Speaking at Davos on 19 January 2026, Ng emphasised that “for many jobs, AI can only do 30-40 per cent of the work now and for the foreseeable future.” This technical reality underpins his broader argument: the challenge is not mass technological unemployment, but rather a widening productivity gap between those who develop AI competency and those who do not. The implication is stark-in a world where AI augments rather than replaces human labour, the person wielding these tools becomes exponentially more valuable than the person without them.

Understanding the Talent Shortage

The urgency behind Ng’s call for skills development is rooted in concrete market dynamics. According to research cited by Ng, demand for AI skills has grown approximately 21 per cent annually since 2019. More dramatically, AI jumped from the 6th most scarce technology skill globally to the 1st in just 18 months. Fifty-one per cent of technology leaders report struggling to find candidates with adequate AI capabilities.

This shortage exists not because AI expertise is inherently rare, but because structured pathways to acquiring it remain underdeveloped. Ng has observed developers reinventing foundational techniques-such as retrieval-augmented generation (RAG) document chunking or agentic AI evaluation methods-that already exist in the literature. These individuals expend weeks on problems that could be solved in days with proper foundational knowledge. The inefficiency is not a failure of intelligence but of education.

The Architecture of Ng’s Approach

Ng’s prescription comprises three interconnected elements: structured learning, practical application, and engagement with research literature. Each addresses a specific gap in how professionals currently approach AI development.

Structured learning provides the conceptual scaffolding necessary to avoid reinventing existing solutions. Ng argues that taking relevant courses-whether through Coursera, his own DeepLearning.AI platform, or other institutions-establishes a foundation in proven approaches and common pitfalls. This is not about shortcuts; rather, it is about building mental models that allow practitioners to make informed decisions about when to adopt existing solutions and when innovation is genuinely warranted.

Hands-on practice translates theory into capability. Ng uses the analogy of aviation: studying aerodynamics for years does not make one a pilot. Similarly, understanding AI principles requires experimentation with actual systems. Modern AI tools and frameworks lower the barrier to entry, allowing practitioners to build projects without starting from scratch. The combination of coursework and building creates a feedback loop where gaps in understanding become apparent through practical challenges.

Engagement with research provides early signals about emerging standards and techniques. Reading academic papers is demanding and less immediately gratifying than building applications, yet it offers a competitive advantage by exposing practitioners to innovations before they become mainstream.

The Broader Theoretical Context

Ng’s perspective aligns with and extends classical economic theories of technological adoption and labour market dynamics. The concept of “skill-biased technological change”-the idea that new technologies increase the relative demand for skilled workers-has been central to labour economics since the 1990s. Economists including David Autor and Frank Levy have documented how computerisation did not eliminate jobs wholesale but rather restructured labour markets, creating premium opportunities for those who could work effectively with new tools whilst displacing those who could not.

What distinguishes Ng’s analysis is its specificity to AI and its emphasis on the speed of adaptation required. Previous technological transitions-from mechanisation to computerisation-unfolded over decades, allowing gradual workforce adjustment. AI adoption is compressing this timeline significantly. The productivity gap Ng identifies is not merely a temporary friction but a structural feature of labour markets in the near term, creating urgent incentives for rapid upskilling.

Ng’s work also reflects insights from organisational learning theory, particularly the distinction between individual capability and organisational capacity. Companies can acquire AI tools readily; what remains scarce is the human expertise to deploy them effectively. This scarcity is not permanent-it reflects a lag between technological availability and educational infrastructure-but it creates a window of opportunity for those who invest in capability development now.

The Nuance on Job Displacement

Importantly, Ng does not claim that AI poses no labour market risks. He acknowledges that certain roles-contact centre positions, translation work, voice acting-face sharper disruption because AI can perform a higher percentage of the requisite tasks. However, he contextualises these as minority cases rather than harbingers of economy-wide displacement.

His framing rejects both technological determinism and complacency. AI will not automatically eliminate most jobs, but neither will workers remain unaffected if they fail to adapt. The outcome depends on human agency: specifically, on whether individuals and institutions invest in building the skills necessary to work alongside AI systems.

Implications for Professional Development

The practical consequence of Ng’s analysis is straightforward: professional development in AI is no longer optional for knowledge workers. The competitive dynamic he describes-where AI-capable workers become more productive and thus more valuable-creates a self-reinforcing cycle. Early adopters of AI skills gain productivity advantages, which translate into career advancement and higher compensation, which in turn incentivises further investment in capability development.

This dynamic also has implications for organisational strategy. Companies that invest in systematic training programmes for their workforce-ensuring broad-based AI literacy rather than concentrating expertise in specialist teams-position themselves to capture productivity gains more rapidly and broadly than competitors relying on external hiring alone.

The Hype-Reality Gap

Ng’s emphasis on “getting through the hype” addresses a specific problem in contemporary AI discourse. Public narratives about AI tend toward extremes: either utopian visions of abundance or dystopian scenarios of mass unemployment. Both narratives, in Ng’s view, obscure the practical reality that AI is a tool requiring human expertise to deploy effectively.

The hype creates two problems. First, it generates unrealistic expectations about what AI can accomplish autonomously, leading organisations to underinvest in the human expertise necessary to realise AI’s potential. Second, it creates anxiety that discourages people from engaging with AI development, paradoxically worsening the talent shortage Ng identifies.

By reframing the challenge as fundamentally one of skills and adaptation rather than technological inevitability, Ng provides both a more accurate assessment and a more actionable roadmap. The future is not predetermined by AI’s capabilities; it will be shaped by how quickly and effectively humans develop the competencies to work with these systems.