Work in the age of AI: From asset-centric to human-centric thinking

The management systems, work structures, and training assumptions we inherited from the industrial era are no longer fit for purpose.

This was the sobering but forward-looking message emerging from last month's inaugural Career Health Summit in Singapore. In a keynote address to gathered policymakers, analysts, business leaders, and HR leaders, Professor Joseph Fuller of Harvard Business School described how, with the rise of generative AI, demographic contractions, and a global productivity slowdown, we are entering a new phase of economic reality — one that demands we abandon the comfort of outdated principles and urgently embrace new frameworks for growth.

“The Taylorist era is coming to an end,” Fuller declared. “Inverting our assumptions about labour and capital is no longer optional — it’s necessary.”

Frederick Winslow Taylor’s scientific management, developed in the early 20th century, treated labour as an abundant, low-cost input and capital as the scarce asset to be optimised. The result? A rigid, hierarchical workplace structure that divided jobs into repeatable routines, prized compliance, and assumed long, linear career paths.

“The goal was to make people as machine-like as possible,” Fuller explained. “Under Taylorism, workers served assets. The more standardised human behaviour, the better.” 

These ideas didn’t just shape factories. They shaped modern job descriptions, recruitment processes, and career paths, but they are now out of date.

Today, the assumptions behind Taylorism have flipped. Assets are plentiful, thanks to cheap computing, ubiquitous cloud tools, and generative AI. Human talent, particularly skilled and adaptable workers, is now the bottleneck. “Capital is no longer scarce. It’s people who are scarce — and it’s getting worse,” Fuller warned.

This "great inversion" redefines strategic priorities. Human capital, once seen as replaceable, is now the most constrained and valuable input in the economy.

The labour market crisis has been brewing for years and we can no longer ignore it. Countries like China and South Korea are facing a 20% decline in the working-age population over the next 25 years. Singapore, while more resilient, also faces a structural slowdown in population growth.

On top of this, labour force participation is falling, especially among prime-age men. Many who could be working aren’t, and employers can no longer assume entry-level talent will show up. And those who do enter the workforce face the hurdle of productivity.

In decades past, productivity gains fueled economic growth. In the post-WW-II era, countries like Germany, France, and Japan leapfrogged the US in productivity by rebuilding modern infrastructure. Later, Singapore, China, and Korea saw major gains by integrating into global trade. But today, productivity is sagging globally. 

Fuller believes generative AI may be the most significant technological development for productivity since the invention of controllable power. In his research with Accenture, Fuller and his team assessed 900 occupations tracked by the US government, analysing how much of each job could be displaced by AI. The result? On average, 41% of work across these roles could be displaced within five years.

“That’s not a forecast. That’s an average across all jobs, 41%,” Fuller emphasised.

Manual roles (like roofing) face relatively low displacement (9%), while high-paying, cognitive, and routine jobs are much more at risk — some over 90%.

Here’s where the disruption deepens. Fuller’s research shows that AI doesn’t just affect core job tasks, it primarily affects supplemental skills, those surrounding the job’s core functions that often go unrecognised in hiring.

This pattern is already visible in the US data. In industries more exposed to AI, the range of skills employers seek has widened, and the rate of job description change is faster. Employers in these sectors are already shifting from narrowly defined role-based hiring to seeking diverse, adaptable skill sets.

And, every level is affected, not just entry-level roles. Using data from seven organisational levels — entry to executive leadership — Fuller emphasised how his research shows that AI is driving both displacement and augmentation of tasks across the board.

“This is not just a training issue for fresh grads,” Fuller noted. “This is about upskilling your entire workforce, leadership included.” Companies will need to provide continuous learning and real-time reskilling opportunities across all job tiers, not just rely on historical training models.

AI’s impact will also fundamentally alter career durability. Workers will have shorter tenures in any one role or skillset. They’ll need to translate existing skills into new contexts, continuously add new capabilities, access dynamic learning systems — not just formal schooling, and understand labour market trends and opportunities in real time.

A compelling example of AI’s productivity impact is in software engineering. In the US, 60% of code written this year has been generated by AI. Tools like GitHub Copilot and Cursor have made senior engineers up to 60% more productive. These engineers are no longer doing the simple manual task of writing code, Fuller said: they know how to spot hallucinations, write prompts, and direct AI effectively.

But meanwhile, wages are falling for junior roles, and job postings for entry-level talent are declining. The real talent war is now for experienced, AI-augmented workers, and companies must both develop and retain them before competitors poach them.

And no one is immune. AI is no longer confined to low-skill or repetitive tasks. Fuller shared how Microsoft recently tested an AI diagnostic tool on 300 historically difficult medical cases. It outperformed physicians in diagnostic accuracy. This doesn’t eliminate the role of physicians, but it does transform it. AI will assist in routine data analysis, freeing doctors to spend more time with patients, which is why most chose medicine in the first place.

In today’s labour market, relying on spot hiring is increasingly unsustainable. It is important to rethink job descriptions to align with evolving tasks, provide access to training across a wider array of skills, work with industry peers and government to ensure national workforce resilience, and abandon rigid Taylorist HR systems that assume static roles.

Companies must upskill and reskill incumbents continuously, because the skills needs of even existing jobs are shifting faster than ever.

Governments have a vital role in supporting the labour market transition like fixing information asymmetries, encouraging employer investment in training, integrating industry-wide efforts for scalable interventions, and modernising education to emphasize soft skills and human work.