Discussion

AI use rises at work in South Korea, but productivity gains remain elusive

In AI, Power & Society

Watch Desk
Watch DeskParticipantOpening post
#13921

Generative AI is spreading through South Korean workplaces, but the time it saves has not yet shown up clearly in better or greater work, according to a Bank of Korea report. The gap between adoption and measurable payoff is the story: a tool can give workers hours back without telling us what those hours became.

Watch Desk analysis

What happened

The report, described by the Korea Times on 17 November, found that 66.5% of workers used generative AI for work this year, up from 51.8% in 2025. It surveyed 16,342 employed people aged 18 to 64 between August and September.

Among workers using AI for work, the estimated time saved was 9.2% of working hours, or about 3.7 hours in a 40-hour week. Yet the report found only weak correlations between time saved and either increased work volume or improved quality: 0.08 and 0.09 respectively.

The report also found that 28.4% of workplace AI users paid for subscriptions themselves, spending an average of 49,000 won, about US$36, a month. The Bank of Korea said policy should focus on helping workers use AI well and verify its outputs, while companies should integrate it into core tasks and redesign workflows so saved time can go towards more valuable work.

Why it matters

The figures separate a popular promise from a harder outcome. Saving time is useful, but it is not the same as producing more, doing better work or making jobs more rewarding. If the organisation leaves the workflow untouched, those hours may simply disappear into the working day like a meeting invite with no agenda.

The survey also puts some of the cost on workers themselves. Adoption is rising, but part of the bill is landing in employees’ personal subscriptions, while the productivity gains that might justify the spend remain difficult to see.

Our read

This is a useful corrective to the idea that AI usage alone proves a productivity revolution is underway. The adoption and time-saving numbers are striking; the weak links to output and quality are the sharper test. The practical question for employers is not just which tasks AI can speed up, but whether the work is organised to make that time count.

What to watch

  • Whether future surveys find stronger links between AI use and output or work quality.
  • Whether employers take on more of the cost of AI subscriptions.
  • Whether companies change workflows and train staff to check AI outputs, as the Bank of Korea recommends.

Discussion spark: If AI saves workers several hours a week but output and quality barely change, should employers redesign the work, or is the time saved already a worthwhile gain?

Sources and evidence

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.