Talent and technology: what GTCI 2017 got right about the future of work

A talent index is an odd place to look for the future of work. It ranks countries, not tasks. Yet the 2017 edition of the GTCI chose technology as its theme, and came back with seven messages that read, with hindsight, like a checklist for the next decade¹. That makes them testable. Some of the evidence has since moved to numbers: employer surveys from the World Economic Forum, a task-level exposure index from the ILO, and an AI chapter from the OECD. This page sets the messages against that evidence and rates each one.

Why technology was the theme

The theme was a bet on a side of a long argument. Detractors say technology impoverishes workers, supporters say it relieves toil and creates new posts, and the 2017 Adecco foreword leaned to the second view, saying the edition analyses how technology can indeed augment human efficiency¹. The Human Capital Leadership Institute’s foreword was darker. Jobs, it said, were being created and destroyed at an unprecedented rate, and the old reassurance that technology merely moves jobs around was “perhaps less of a fallacy” than it used to be¹.

Two ideas held the edition together. Technology boosts growth by raising productivity and lowering costs, but also brings, temporarily at least, potential skills shortages as new types of human expertise become necessary¹. And the new expertise is not only technical: the edition highlights “softer” talents such as creativity, adaptability and the ability to work in teams¹.

The index tied the theme to its model. The 2017 model added an Employability sub-pillar under Vocational and Technical Skills, measuring skill gaps and labour market mismatches in each country¹. The full table of that edition’s results is on the 2017 results page. This page follows the argument instead.

Seven messages, nine years later

Technical skills PLUS social/project competence is the new talent profile.¹

The 2017 report listed seven key messages¹. The table gives each one in short form, the later evidence, and the editor’s reading of how it has aged. The verdicts are judgements, grounded in the cited lines and no more certain than the evidence.

The first message has aged best. “Think beyond automation” asked decision-makers to see the whole system moving, not only robots replacing humans: organisations, careers, and education and employment systems still founded on what the report called a fast-fading twentieth-century factory model¹. The later numbers do show jobs being reshaped faster than they are removed. The report did also expect many unskilled jobs to be automated, through robots and algorithms empowered by big data¹. That half of the forecast is still open: the OECD sees low- and middle-skilled jobs, among them construction, farming and production, as the most at risk⁶.

The third message is the most useful for readers deciding what to learn. It has been confirmed from three sides. Employers rank analytical thinking first³. The OECD finds human skills gaining more than specialised AI skills⁶. And the GTCI itself, in 2025, relabelled its high-level skills pillar Generalist Adaptive Skills to reflect soft cooperative skills, adaptability and interdisciplinary problem-solving⁷.

Jobs: created, displaced, churned

The World Economic Forum’s Future of Jobs survey is the most cited forecast of the question. The 2025 edition draws on over 1,000 employers representing more than 14 million workers in 22 industry clusters and 55 economies, and looks across 2025 to 2030³. Its central estimate is 170 million jobs created and 92 million displaced, a churn of 22% of the 1.2 billion formal jobs in its dataset and a net increase of 7%, or 78 million jobs³.

The 2023 edition looked five years ahead and saw the opposite sign: 83 million jobs lost and 69 million created, a churn of 23% of the 673 million employees in its dataset, a net reduction of 14 million jobs, or 2%⁴.

Worked example (our arithmetic). The two editions use different samples and horizons, so the counts cannot be set against each other. Dividing each edition’s counts by its own dataset puts them on the same footing^3,4.

Two things follow. Churn, the share of jobs that change hands, is almost the same in both, 23% and 22%^3,4. The balance between creation and destruction flipped within two years. A fair reading is that employers agree on how much turbulence is coming and disagree about its net direction, which is exactly where a survey of employers is weakest. Treat the sign as a mood, not a measurement.

Which jobs? The 2025 survey expects frontline roles to see the largest growth in absolute numbers, among them farmworkers, delivery drivers and construction workers, with care and education jobs also growing³. Technology roles grow fastest in percentage terms: big data specialists, fintech engineers, AI and machine learning specialists and software developers³. Clerical and secretarial workers see the largest decline in absolute numbers, and postal clerks, bank tellers and data entry clerks the fastest percentage declines³. Ageing populations push up healthcare jobs, while growing working-age populations push up education jobs³.

Skills: the 2017 profile, priced

The 2017 report warned of temporary skills shortages¹. Eight years later the World Economic Forum puts a number on them. Workers can expect 39% of their existing skills to be transformed or outdated over 2025 to 2030, down from 44% in the 2023 edition and a high of 57% in 2020³. One possible reason: half of workers (50%) had completed training, reskilling or upskilling, against 41% in the 2023 edition³.

Worked example (our arithmetic). Employers’ 2025 picture of a workforce of 100 people: 59 need training by 2030. Of these, 29 can be upskilled in their current roles, 19 upskilled and redeployed within their organisation, and 11 are unlikely to receive the training they need³. The three groups add up to the 59: 29 + 19 + 11 = 59.

Skill gaps are the biggest barrier to business transformation, cited by 63% of employers, and 85% plan to prioritise upskilling³. In the same survey 70% expect to hire staff with new skills, 50% plan to move staff from declining to growing roles and 40% plan to reduce staff whose skills become less relevant³. The last number is the uncomfortable one: for roughly four in ten employers, the answer to a skills gap includes letting people go³.

On the skills themselves, AI and big data top the list of fastest-growing skills, followed by networks and cybersecurity and technology literacy³. Creative thinking, resilience, flexibility and agility, and curiosity and lifelong learning rise alongside³. That is the 2017 message about technical plus social skills, rendered as an employer shopping list. The 2020 edition of the GTCI added a note on how the pairing works in practice: re-skilling should develop fusion skills so that humans and machines interact in hybrid activities, and human-AI teams could be more productive than either alone².

Who is exposed

More than replacing jobs, AI is changing them.⁶

Exposure is not loss, and the sources are careful about it. The ILO’s 2025 index puts one in four workers in an occupation with some generative AI exposure, with 3.3% of employment in the highest category⁵. The authors stress that exposure does not imply the immediate automation of an entire occupation, only the potential for a large share of its tasks to be performed with the technology⁵.

The shape of exposure is lopsided. Overall exposure is 11% of employment in the low-income group, and 34% in the high-income group⁵. Clerical occupations have the highest exposure, and exposed roles such as clerical, financial and customer service jobs are concentrated in higher-income countries⁵. Women are more exposed in the highest category: 4.7% of female employment against 2.4% of male employment⁵.

The OECD’s 2023 work looks at the same question from earlier AI. It found the impact on job levels limited so far, with adoption concentrated in the largest firms, many of them still experimenting and adjusting through slowed hiring, voluntary quits and retirement rather than dismissals⁶. Occupations at the highest risk of automation account for about 27% of employment, and three in five workers worry about losing their jobs entirely to AI within ten years⁶. High-skill occupations have the lowest risk of automation⁶.

An editor’s note on a mismatch that readers will notice. The OECD’s list of most-at-risk jobs, built in 2023, runs to construction, farming and production⁶. The ILO’s list, built in 2025 for generative AI, puts clerical and financial work first⁵. The two are not in contradiction. They measure different technologies in different years, and the ILO itself notes that some strongly digitised occupations have gained exposure as the tools improve⁵. The safe statement is that nobody’s list of exposed jobs is stable.

Where countries stand

The 2017 edition assessed 118 countries and concluded that nine were particularly well positioned in talent readiness for technology: Switzerland, Singapore, the United Kingdom, Denmark, the Netherlands, Ireland, Canada, New Zealand and the United Arab Emirates¹. In the same 2017 edition Switzerland ranked first and Singapore second¹.

The 2020 edition, on artificial intelligence, found the gap between talent champions and the rest widening, and a similar gap in AI talent, which it called scarce and unequally distributed². It also found that some emerging countries, including China, Costa Rica and Malaysia, could become talent champions in their regions². By 2025 the index measured AI directly, adding an indicator for AI skills migration⁷. The full table of every edition’s top three is in every edition from 2013 to 2023, and what the index measures shows where technology sits in the model.

What the evidence cannot say

Four limits keep this reading honest. Employer surveys record expectations, and expectations moved from a net loss to a net gain in two years^3,4. Exposure indexes record how much of an occupation’s tasks a tool could perform, not whether employers will use it⁵. Early OECD evidence concerns a technology still at an early stage of adoption⁶. And the GTCI itself is an index of national conditions for talent, not a forecast of jobs. The technology section collects the pages on this topic.

Editor’s verdict

On the evidence of this page, the 2017 diagnosis has aged well and its alarm has not¹. Think-beyond-automation, the new technical-plus-social profile and the need for education and employment systems to adapt all find support in 2023 to 2025 sources. What the OECD’s evidence does not show is a net loss of jobs⁶. What the World Economic Forum does show is churn on the scale of a fifth of all jobs in five years, and a skills gap that employers rank as their biggest obstacle³. For anyone reading a future-of-work headline, the useful questions are the ones this page asks of each source: which edition, which sample, which horizon and whether the number is a forecast, an exposure or a count.