56% of workers in Cyprus hold jobs highly exposed to AI: what it means for offices

What happened

According to the Cyprus Mail of 1 November 2026, the European Commission report was published the week before and found no evidence so far of net job creation or destruction from AI in the EU. It does record a hiring slowdown since 2022 in occupations with high AI use.

The same Cyprus figures appear in the IMF annex on the labour market, part of the 2026 Article IV report concluded on 17 June 2026, as presented by Politis on 10 October. Of the 56% in highly exposed jobs, 33% are in roles where AI mainly helps the worker and 23.4% in roles where it could replace them. In the private sector that second share is about 27%, in the public sector about 10%. Fewer than 10% of companies have integrated AI technologies.

The most exposed occupations are business administration, clerical work and ICT, and the most exposed industries ICT, finance and professional services. Andreas Matsas of the Sek union said there is no evidence of mass layoffs, but that accountancy and law firms may see cuts. Philokypros Rousounides of Keve said it is too early to tell and that the right training will help.

“if we manage the matter correctly, we will not have job losses here”

What it means for an office in Cyprus

Exposure does not mean layoffs. It means that much of the work in an office, such as email, documents, data entry and reports, can be done faster with AI. The real question for each business is whether AI is placed next to its people or simply used to cut posts without a plan. Offices that organise their tasks, train their staff and keep a person on the final check move into the 33% where AI helps.

With fewer than one in ten businesses already using AI, those who start now gain a lead over their competitors. There is a legal side too: since 2 February 2025, Article 4 of the EU AI Act requires anyone using AI systems to ensure sufficient AI literacy among their staff. This is not legal advice.

What to do

• List the office's weekly tasks and mark which are repetitive work with documents, email or data
• Pick one or two tasks where AI helps staff, such as drafts, summaries or invoice entry, and trial them for four weeks with human review
• Send managers to subsidised training, such as the two-day Keve AI course in Nicosia on 11 and 12 November, funded by the HRDA for eligible participants
• Write a short AI use policy: which tools, which data never goes into them because of GDPR, and who checks the output
• Talk openly with staff about which tasks change and where the time saved will go