IMF: AI-exposed jobs now seek AI skills twice as often; coding demand dips 10%

Jobs that are most exposed to artificial intelligence now ask for AI skills about twice as often as jobs that are less exposed, according to research by the International Monetary Fund presented on Monday, October 12, 2026, at the IMF and World Bank annual meetings in Bangkok. The analysis covered about 700 million job advertisements. The IMF reads the result as employers changing the skills they want from workers rather than simply replacing people with machines. But the research also found a cost: hiring demand for some highly exposed roles, such as programming, has fallen by about 10 per cent compared with less-exposed occupations. An IMF official for the Asia-Pacific region said that decline has since levelled off, even as AI tools have become more capable.

The debate over AI and work is often framed as a simple question of whether machines will take jobs. The IMF's job-advert data give a more detailed answer. In the occupations where AI can do the most, employers are not just cutting posts; they are rewriting job descriptions to ask for people who can use AI well. For workers, that turns AI from a threat into a skill that can be learned and shown on a CV.

The 10 per cent fall in relative demand for programming roles is a warning for India, where many graduates aim for software and coding jobs. The IMF's finding that the drop has stabilised offers some reassurance, but it suggests that a coding degree alone is no longer enough. Employers in exposed fields want the three kinds of AI skill the IMF identified on top of core technical ability.

The third group, integrating AI with a company's own data, is relevant for IT services workers who build systems for clients. Connecting AI models to an organisation's records, documents and processes is what turns a general tool into something a business can use day to day.

The concern for entry-level workers is also real. Speaking around the meetings, IMF Managing Director Kristalina Georgieva highlighted pressure on young people starting their careers, as employers in exposed sectors find they can automate tasks that junior staff used to do. That makes early training in AI tools a sensible step for students.

Short-term: Students and workers in exposed fields, including software, can expect more job adverts to list AI skills, from basic literacy to prompt engineering and data integration, as requirements.

Long-term: Workers who add AI skills to their core expertise are likely to find more openings, while those in routine roles that AI can do may face slower hiring. Training institutions and employers that teach these skills will shape who benefits.

Who is affected: Software developers and IT services workers, engineering and computer science students, entry-level job seekers, training providers and employers redesigning roles around AI.

Questions and Answers