Commentary: Singapore should not focus only on productivity gains from AI
AI is critical, but what is more important for long-term gains is how it is learned and adopted, says Manu Kapur of the Singapore-ETH Centre.
New: You can now listen to articles.
This audio is generated by an AI tool.
Read a summary of this article on FAST.
Get bite-sized news via a new
cards interface. Give it a try.
Click here to return to FAST
Tap here to return to FAST
FAST
SINGAPORE: In a world of tighter labour markets and rising costs, artificial intelligence is being positioned not as a nice-to-have, but as a critical lever for economic competitiveness and national advancement.
Singapore has established a national council to steer the AI agenda, lower the costs for businesses to adopt AI, and support individuals who sign up for selected AI courses.
However, there is a critical risk lurking inside every AI strategy, whether for a country, company or individual: What works for productivity may not necessarily work for learning and building capability.
THE AI PRODUCTIVITY TRAP
AI’s immediate promise is obvious. It drafts emails, summarises reports, generates slide decks, proposes marketing copy and even recommends decisions. Output goes up, response times fall, costs drop and KPIs are met.
Guess Word
Crack the word, one row at a time
Buzzword
Create words using the given letters
Mini Sudoku
Tiny puzzle, mighty brain teaser
Mini Crossword
Small grid, big challenge
Word Search
Spot as many words as you can
When tasks become easier thanks to technology, people perform better in the moment but acquire less understanding that can be transferred to new situations. A recent study of more than 26,000 Chinese students over 30 months showed that although the use of AI improved homework performance and reduced completion times, it hurt students’ learning and exam scores.
This should matter to businesses that want to invest in their workers. Consider this example: A junior lawyer uses an AI copilot to draft a contract in minutes for the senior partner to approve. On the surface, AI has helped the junior associate accelerate their productivity.
However, for the firm, the task is not complete without the senior partner’s go-ahead. The partner carries the burden of accountability because of their domain knowledge and capability.
If AI does the heavy lifting, junior lawyers may not learn why an indemnity clause matters, when confidentiality carve-outs are risky, or when a data-protection obligation should be escalated.
All of these require domain knowledge and judgment. This expertise can only be built when juniors practise drafting contracts, receive feedback, confront what they do not understand, and then try again.
This in turn helps the firm derive even greater gains from AI in the longer run, where employees not only use AI but critically evaluate AI output. The same argument applies to sectors where junior staff must be trained, from consulting, accounting and finance to medicine, policy-making and education.
A TWO-TRACK ADOPTION PLAN
When companies send their employees for AI training, employees must learn how to diagnose problems, test assumptions, verify sources and transfer what they learn to new situations. For most businesses, however, the fastest way to integrate AI is to deploy it wherever their work processes allow.
A more durable approach for a business to embrace AI is to run a two-track adoption plan. In the first track – the performance track – businesses could deploy AI where speed and consistency matter, with proper governance and monitoring.
In the second track – the learning track – businesses should deliberately create settings where teams of employees cannot outsource thinking. In these settings, employees would have to experiment with different solutions, critique AI outputs, run post-mortems on failures and develop internal playbooks.
Businesses which run only Track A adoption will undoubtedly get quick wins but they will hit a ceiling when the environment changes. Businesses which also adopt Track B will build organisations that will keep adapting, growing and winning.
This is why business leaders must be cautious about AI layoffs. If firms use AI only to reduce headcount, they may enjoy short-term savings but create a capability debt. They will have fewer people who deeply understand the workflow, exceptions, customers and risks.
Instead of focusing on “How many tasks can AI replace?”, firms would do well to ask: “Which human capabilities must we now build because AI has changed the task?”
Some jobs will evolve, some will disappear and new ones will emerge. But the firms that win will be those that redesign work so people move into higher-value roles as AI takes over routine work.
MEASURING AI GAINS
For policymakers, the same logic applies on a national scale. Singapore’s National AI Council is promising because it can coordinate action across sectors. But we should not measure AI success only by adoption rates, number of tools deployed or near-term productivity gains.
A better scorecard would ask four questions:
- Productivity: Are firms saving time, reducing costs and improving service quality?
- Reliability: Are AI-enabled processes reducing errors, or merely producing mistakes faster?
- Capability: Are workers getting better at diagnosing problems, challenging AI outputs and explaining decisions? This should be assessed not only while AI is available, but also in situations where workers must reason independently.
- Innovation and resilience: Are organisations using AI to create new services, improve processes and adapt to new problems, instead of simply automating yesterday’s workflows?
Singapore has signalled it is serious about AI execution. The next step is to be equally serious about what execution alone cannot guarantee – learning.
If we treat AI only as a productivity machine, we risk scaling unproductive success. Such success might yield impressive output today, but lead to brittle and unadaptive capability tomorrow.
If instead we design AI adoption around both performance and learning, we can build a workforce and an economy that can keep adapting as the technology inevitably changes again.
Manu Kapur is Director of the Singapore-ETH Centre and Professor of Learning Sciences & Higher Education, ETH Zurich, Switzerland. He is author of Productive Failure: Unlocking Deeper Learning Through The Science Of Failing.
Source: CNA/zw(el)
Sign up for our newsletters
Get the CNA app
Stay updated with notifications for breaking news and our best stories
Get WhatsApp alerts
Join our channel for the top reads for the day on your preferred chat app
Get bite-sized news via a new
cards interface. Give it a try.
Click here to return to FAST
Tap here to return to FAST
FAST
About The Author
