Is It Better to Specialize or Stay a Generalist in the Age of AI?

By Emil Bjerg

The question used to be about preference – but AI is giving it a new urgency.

The classic high-earning career path had a clear logic: pick a lane, go deep, become indispensable. Spend your twenties mastering something specific – financial modeling, UX research, or tax law – and let your expertise compound over time. It worked brilliantly in a world where domain knowledge depreciated slowly and roles stayed stable for decades.

With AI, that world is dying. AI is compressing the learning curve in dozens of specialized domains, automating the most routine layers of expert work, and creating entirely new roles that didn’t exist three years ago. The question of whether to specialize or stay a generalist has become one of the most consequential career decisions a young professional can make – and the answer is considerably more nuanced than either camp usually admits.

What the data says

The World Economic Forum’s Future of Jobs Report 2025 – drawn from over 1,000 major employers representing 14 million workers – found that nearly 40% of job skills are expected to change by 2030, with 92 million roles displaced and 170 million new ones created. The net figure sounds reassuring – +78 million jobs – but the churn underneath it is unprecedented.

Analytical thinking tops the list as a core requirement for 69% of employers. Resilience, flexibility, and agility come second at 67%.

It’s therefore worth understanding what employers say they want. Analytical thinking tops the list as a core requirement for 69% of employers. Resilience, flexibility, and agility come second at 67%. Creative thinking, curiosity and lifelong learning, and leadership round out the top skills. AI and big data literacy, meanwhile, are projected to see the fastest growth in demand of any skill cluster across almost every sector.

In other words, a mix of hard technical capability (AI and data literacy) and classically “generalist” traits (adaptability, creative thinking, curiosity).

Meanwhile, Upwork’s 2025 In-Demand Skills report registered a 220% year-over-year growth in specialized AI roles with those positions commanding up to 22% higher hourly rates than traditional AI work. Specialization – at least in the right domains – still pays a premium.

Why the specialist model is under pressure

The specialist’s traditional moat was access to knowledge. A tax attorney, a radiologist, or a data engineer’s value come partly from knowing things most people couldn’t easily learn. AI is quietly eroding that moat by commoditizing the most routine layers of their work.

As Gartner research highlighted in HBR, AI is simultaneously creating excess capacity in legacy roles and acute shortages in AI-adjacent skills. A global survey of C-suite executives found nine out of ten reporting workforce overcapacity of up to 20% in traditional roles alongside skill shortfalls in AI-critical areas. The specialist who learned one stack, one methodology, one domain – and stopped there – is most exposed.

Why the generalist model is also insufficient

The counterargument – that generalists will thrive because they’re adaptable and AI does the specialist work for them – is also not the full truth: depth still matters enormously.

A generalist with no specific area of expertise risks being easily replaced by a sufficiently capable AI assistant and someone with more domain-specific judgment than they have. The domains where contextual understanding, ethical reasoning, and genuine mastery are irreplaceable – high-stakes medicine, legal strategy, structural engineering, complex research design – still reward deep expertise. A radiologist using AI to pre-screen images and focus cognitive effort on the difficult cases is more valuable than before, not less.

David Epstein’s Range: Why Generalists Triumph in a Specialized World made the case that breadth and diverse experience produce more creative, adaptable problem-solvers – particularly in what he calls “wicked” environments where rules are unclear. His argument remains valid. But even Epstein’s generalists typically have at least one area of genuine depth. The safer way to be a generalist is to be able to apply transferable skills – creativity, communication, adaptability – across contexts, but with a domain expertise that gives you somewhere to stand.

The T-shape revisited

A useful framework is McKinsey’s T-shape – the idea that the most effective professionals combine broad lateral skills across domains with deep vertical expertise in at least one area. The horizontal bar is range; the vertical bar is credibility. McKinsey has used this as an internal recruiting concept since the 1980s, and during the current AI-hiring shift, the firm has foregrounded it again.

The firm is now actively recruiting liberal arts graduates – candidates who, as CEO Bob Sternfels put it, bring “truly novel” ways of thinking that can make “discontinuous leaps” in logic that AI models cannot. As reported by the Financial Times and Fortune, McKinsey is simultaneously requiring junior candidates to use an internal AI assistant during interviews – demanding both adaptability and AI fluency in the same breath.

Three practical orientations for early-career professionals

1) Specialize in something AI cannot easily compress

The domains that remain robust are those requiring contextual judgment, relational trust, or embodied knowledge. Strategic narrative, clinical assessment, legal argumentation, design sensibility – these still require a human holding the wheel. Technical specializations that are essentially pattern-matching at scale (routine coding, standard data analysis, template legal drafting) are higher-risk over a five-year horizon.

2) Make AI fluency part of your core competency

The WEF’s 2025 data shows the proportion of workers with AI skills has increased by at least 100% across all sectors since 2016 – but demand is outpacing supply. Knowing how to work with AI – directing it, auditing its outputs, identifying where it fails – is increasingly the table stakes for any professional role, specialist or generalist.

3) Treat your career as a sequence of deliberate moves rather than a single bet

The old model of “specialize once, then coast” is about to be a thing of the past. The 22-year education followed by 40 years of stable application no longer maps onto reality. The professionals with the most resilient careers will be those who can reskill into adjacent domains while AI reshapes their current one – which requires building the habit of learning, not just a body of knowledge.

Beyond the specialist versus generalist debate

In the foreseeable future, the best packaging is a mix of domain expertise and generalist skills.

The professionals who will struggle are those at the extremes: the narrow specialist who learned one thing and stopped learning, and the pure generalist with no real depth who assumed breadth alone would protect them. The ones who will thrive are those who can hold both – a genuine area of mastery, and the intellectual flexibility to apply it across shifting contexts with AI as a collaborator rather than a competitor.

That’s more continuous – and perhaps harder – than picking one lane. It also appears to be the only approach that makes sense in a market where the lanes keep moving.