To boldly go: redefining jobs and expertise in the AI era

WE FIND it unproductive to be pulled between utopian visions of artificial intelligence (AI)-driven efficiency and dystopian fears of mass unemployment.

What we observe is more interesting: AI is revealing what was always most valuable about human work – and it isn’t what most of us were trained to sell.

AI can now draft reports, write code and analyse data – the bedrock of modern jobs. If your value comes from recalling and synthesising information, that value is falling fast.

But something else is becoming visible.

As routine cognitive work gets automated, the work that remains is decisively human: the judgment call made with incomplete information, the emotional intelligence to navigate a difficult client, the ability to know when the model is confidently wrong – and the willingness to be accountable for the outcome.

AI can recommend. It cannot sign off. It cannot stake its reputation. Every organisation, every market, every society requires someone who owns the consequences. That someone is human – and that role is becoming more valuable, not less, as AI handles everything else.

The partnership problem

This is fundamentally good news, but it comes with a catch.

The tasks being automated – reviewing contracts, building financial models, managing projects end to end – aren’t just tasks. They are how people have traditionally developed the judgment that can’t be automated.

A lawyer learns to spot risk by reviewing hundreds of contracts. An analyst develops intuition by building models that break. A leader learns execution by managing the messy middle of projects. When AI handles these, the training ground narrows – precisely when we need what it produces most.

So the question isn’t whether humans remain relevant; they do, more than ever. The question is how we develop the capabilities that make us relevant, when the old pathways to building them are changing.

This is a design problem, not a crisis. And some organisations are already working on it.

How work is being redesigned

Cognitive flight simulators: Aviation solved expertise development without real crashes. The same logic applies here. Protected environments where professionals use AI to run complex scenarios: What happens to our supply chain if oil hits US$200? Redesign this product for emerging markets. Failure is cheap, learning is real, and the point is sharpening judgment – not efficiency.

Adversarial collaboration: For a century, professional work meant creation – writing the first draft, building the first model. AI now produces capable first drafts. The emerging shift is from “creators” to “red teamers”: AI generates and humans rigorously stress-test, hunting errors, injecting context AI lacks and challenging assumptions. Firms that use AI only for speed will plateau. Firms whose people constantly interrogate AI output will build sharper thinkers.

Portfolios over credentials: A degree shows programme completion; a portfolio shows thinking and delivery. Forward-looking organisations are beginning to test judgment through work samples and trial projects rather than filtering candidates through resumes. This matters because AI levels the playing field on formal knowledge – what differentiates people now is demonstrated capability.

Career development reimagined: The old model – earn by doing routine tasks, progress by osmosis – assumed routine tasks would always exist. New approaches pair professionals directly with mentors on AI-augmented strategic projects, where learning comes from observing and debating judgment rather than executing rote work. Some teams are creating “judgment councils” where cross-level groups debate AI-generated recommendations, making senior thinking visible to junior staff.

What this means for individuals

Organisational change moves at its own pace. Individuals don’t have to wait.

AI has lowered the barriers to demonstrating capability. An analyst can build a market assessment, a designer can prototype a product, an engineer can ship a working app – without institutional permission.

The professionals gaining ground are those who have already shown they can orchestrate AI to solve real problems, building a body of work that speaks for itself.

If your daily work is throughput – processing, summarising, generating routine output – it’s worth asking how long that remains yours to do.

But if your work centres on judgment, on the difficult conversation, on the creative leap that requires understanding what the numbers don’t say, on unknown unknowns – that’s where value is concentrating.

Leaning into the discomfort of those tasks, rather than away from it, is how capability compounds.

The reinforcing dynamic matters. Individuals building portfolios create pressure on organisations to value them. Organisations creating new development pathways give people opportunities to build judgment. Neither alone changes the system, but together, they can.

Singapore’s position

With 77 per cent of its workforce in AI-exposed roles – well above the global 60 per cent average – Singapore will encounter this transformation more quickly and intensely than most economies.

That’s a reason for urgency, but also an advantage: the models we develop for human-AI partnership become exportable expertise. Our fintech regulatory sandbox already demonstrates how to enable experimentation without losing control. The opportunity now is to extend that instinct into how we develop talent and structure work.

The shift starts concretely – a hiring decision that weights capability over pedigree; a workflow redesign that builds judgment instead of just efficiency; and a learning environment where failure is safe.

If we use AI solely to make work easier, we risk making ourselves less capable. The better path is to use AI to make work more challenging – to simulate complex realities, demand higher judgment, and develop the human capabilities that define what AI cannot do.

The technology is clarifying what matters most about human work. Whether we build around that clarity is the choice in front of us – organisations and individuals together.

Taimur Baig is chief economist at DBS. Leslie Teo is senior director at AI Singapore and a member of the United Nations Independent International Scientific Panel on AI.

The writers used several large language models to co-author this editorial – summarising hundreds of papers, building an economic model and drafting various versions. In their view, the AI was tireless and capable, yet utterly lacking in judgment. Read the companion piece here.