Cyprus may be worrying about the wrong labour crisis. The familiar fear is that artificial intelligence will leave people without jobs. The more immediate danger is that the island will have jobs without enough people to fill them. 

The contradiction is already visible. Some 44.2 per cent of Cypriots used generative AI in 2025, placing the country well above the EU average. Among younger people, usage reached 76.5 per cent. However, only 9 per cent of Cypriot businesses used AI, compared with 20 per cent across the EU and 42 per cent in Denmark, according to Eurostat data and its business figures. 

Cypriots have AI in their pockets. Their employers have barely put it to work. 

That gap is appearing just as the workforce begins to feel the weight of demographics. Cyprus now has more people aged 65 and over than children below 15, while its fertility rate remains at 1.4 children per woman, according to the latest Cyprus figures. 

Unemployment stood at just 4 per cent in the first quarter of 2026, but employers were trying to fill almost 14,000 positions, particularly in construction, retail, tourism and food services. Meanwhile, the Human Resource Development Authority (Anad) expects Cyprus to need around 11,700 recruits every year until 2032. More than half will not fill new jobs but replace people permanently leaving work, according to the labour survey, vacancy figures and Anad forecast. 

The idea of AI arriving to steal Cyprus’ jobs therefore misses half the story. It is arriving in an economy that may soon struggle to find the workers needed to keep hotels open, finish construction projects, care for an older population and run thousands of small businesses. 

The latest international evidence offers little support for an imminent employment apocalypse. On Thursday, July 23, Adecco chief executive Denis Machuel told Reuters that some companies were using AI to justify cuts actually caused by weak performance, restructuring or other problems. Three and a half years after ChatGPT’s launch, employment across the Organisation for Economic Co-operation and Development (OECD) remains close to record levels, while Adecco’s new whitepaper estimates that 1.9 million AI-related jobs were created between 2022 and 2025. AI is changing tasks faster than it is eliminating entire occupations, Reuters reported. 

Australia has gone looking for evidence of the damage and found relatively little. Its government published the country’s first dedicated AI employment assessment on July 8 and found no broad upheaval, no collapse in youth employment and no unusual acceleration in workers moving between occupations.  

Some highly exposed roles were growing more slowly, but software development employment had expanded strongly despite being among the occupations most exposed to AI. Exposure, in other words, can generate demand as well as displacement, according to the Australian review. 

Britain has found something equally revealing. AI use among businesses with at least 10 employees almost tripled between late 2023 and June 2026, rising from around 12 per cent to 35 per cent. But only one in ten adopting businesses said it was using the technology extensively. Most reported no change in staff numbers, while training existing employees was considerably more common than recruiting teams of AI specialists, according to the UK data. 

The technology may be spreading rapidly, but the transformation inside companies is much slower. Businesses are still working out which problems AI can solve, who should use it and how it fits into existing jobs. 

Japan shows why that delay matters. It has one of the world’s tightest labour markets, with unemployment at 2.5 per cent and serious shortages among service companies and smaller businesses. Nevertheless, only 8.4 per cent of Japanese workers report using AI at work. Older and non-regular employees are even less likely to use it or receive the training needed to benefit from it, according to OECD research. 

South Korea has moved a little further. More than a quarter of smaller companies that had struggled to recruit said generative AI was helping them cope with the shortage. At the same time, 95.5 per cent of AI-adopting firms reported no reduction in their team or department workforce. More than half said AI had replaced individual tasks within jobs. 

That may be the clearest description of what is happening. Tasks are being replaced more often than teams. The Korean findings also showed that AI increased the need for communication, data interpretation and higher-level skills, rather than making human ability irrelevant, according to the Korean evidence. 

European company data point in the same direction. A study of more than 12,000 firms found that adopting AI increased labour productivity by an average of 4 per cent without reducing employment in the short term. Workers at adopting companies also earned more. The gains were strongest where businesses invested not only in AI but also in software, reliable data and employee training, the firm study found. 

The threat to young workers is more complicated. Employers increasingly expect junior applicants to arrive with judgement, creativity and leadership skills that were once developed through years of routine work. But a newly published analysis of 243 million hires and 407 million job advertisements across the US, Britain, Canada and Australia challenges claims that AI is principally responsible for the decline in entry-level recruitment. 

Once researchers separated AI exposure from the effects of working from home, remote work remained a strong predictor of weaker junior hiring while the AI effect largely disappeared. The authors argue that mentoring and informal workplace learning were disrupted before generative AI became widely used, according to the new research. The career ladder may indeed be breaking, but AI may not be the only hand holding the hammer. 

The most surprising warning comes from Kenya. In a five-month experiment, 640 entrepreneurs received either conventional business guidance or a GPT-4 adviser through WhatsApp. AI produced no significant improvement on average. Stronger businesses gained more than 15 per cent, while weaker ones performed around 8 per cent worse. 

The AI supplied broadly similar advice to both groups. The difference was in what the entrepreneurs chose to follow and how they applied it. The same tool widened the performance gap, according to the Kenyan experiment. 

That finding should matter to Cyprus. Buying access to AI is easy. Knowing where it is useful, questioning its answers and turning them into sensible business decisions require experience and training. Without them, the technology can make a weak process faster rather than make it better. 

Nor does using AI mean becoming an AI engineer. Research published by the International Labour Organisation (ILO) last week found that specialist AI and machine-learning abilities represented only around 1 per cent of the skills requested in online vacancies in Brazil, Egypt, Jordan and the United Arab Emirates. Employers were more consistently looking for people who could combine digital ability with communication, teamwork, analysis and judgement, according to the ILO findings. 

That is a more realistic target for Cyprus. The island does not need every hotel employee, accountant, estate agent or construction manager to learn how to build an AI model. It needs them to understand how the tools can remove paperwork and leave more time for the work requiring experience, responsibility or human contact. 

A chatbot cannot pour concrete, prepare a hotel room or hold the hand of an elderly patient. It can help produce estimates, schedule staff, translate instructions, prepare invoices, check documents and answer repetitive customer questions. In a country short of workers, saving part of a job may be more valuable than eliminating one. 

AI will not remove Cyprus’ need for migration either. One in four residents is already a non-national, while 94 per cent of immigrants arriving in 2024 were foreign citizens, according to European figures. The island will continue to need people from abroad, stronger vocational education and better ways of keeping young Cypriots in the country. What AI can offer is time.