A solid mix of foundational, ICT and complementary skills help empower individuals to navigate an increasingly digital world, participate in the global economy and compete in job markets.
Skills in the AI age
In such an AI and data-driven digital landscape, individuals need a wide range of skills to thrive. These skills can be considered as follows: i) foundational skills that enable all people to participate in a digital economy and society; ii) some basic and more advanced ICT skills that help workers meet evolving job requirements and fully participate in the digital age; iii) complementary skills such as teamwork, autonomy, problem solving, creative thinking, communication, collaboration and emotional intelligence that enable high-performance work practices and a strong ability to continue learning (OECD, 2024[19]).
Foundational skills such as in science, numeracy and reading are essential for everyday activities (OECD, 2019[20]; OECD, 2024[21]). Evidence from the OECD Survey of Adult Skills (PIAAC) shows that high levels of proficiency in literacy and numeracy are associated with better proficiency in problem solving in digital environments. Low levels of proficiency in literacy, and particularly in numeracy, may be significant barriers to using ICT applications to manage information effectively (OECD, 2013[22]).
In addition to foundational skills, individuals need a range of ICT skills that allow them to learn, work and undertake daily activities. Basic ICT skills comprise a minimum level of computer skills, communication and information search skills, and proficiency in using productivity software such as operational and information navigation. Such skills enable access to information, facilitate communication, support civic participation, and enhance employability and well-being (Burns and Gottschalk, 2019[23]); (OECD, 2022[24]).
More advanced ICT skills (e.g. software programming, AI skills) strengthen the ability to cope with change (e.g. in work organisation) and keep learning (e.g. new programming languages). They also promote creativity and innovation and offer numerous career opportunities. “AI skills” refer to the abilities and knowledge required to develop or undertake an advanced use of AI systems. Some key AI skills include programming, understanding and implementing machine-learning algorithms; data mining and data processing; deep learning; natural language processing; and understanding computer vision concepts, such as image processing and object detection.
When combined with a minimum level of ICT skills, complementary skills can significantly enhance an individual’s effectiveness and productivity in digital environments. These are valuable assets across sectors and increasingly sought in today’s technology-driven world that enables high-performance work practices in the digital age (Samek, Squicciarini and Cammeraat, 2021[25]; Borgonovi et al., 2023[26]). The OECD Skills for 2030 project identifies two types of complementary skills: “meta-cognitive skills” include critical thinking, creative thinking, learning-to-learn and self-regulation, while “social and emotional skills” include empathy, self-efficacy, responsibility and collaboration (OECD, 2019[27]).
Despite increased AI adoption by firms, workers with AI skills represent a small share of the overall workforce (about 1%). While workers with AI skills are relatively rare, their share in the workforce has almost tripled from less than a decade ago (Green and Lamby, 2023[28]). Countries are competing for this small pool of highly skilled AI workers (OECD, 2024[29]). Based on live data from the OECD’s AI Policy Observatory, Figure 4 shows that countries such as Luxembourg, Switzerland, Ireland and Germany attracted more AI talent1 than they lost in 2024. In contrast, other economies (e.g. Israel, India and Hungary) saw a net outflow of AI talent from their borders.
The net outflow of AI talent might indicate an AI “brain drain”, where highly specialised workers move to another country for improved employment opportunities. India is noteworthy for its considerable AI talent growth rate in recent years. This rate is ahead of the United States, United Kingdom and Canada, as measured by the year-over-year change in LinkedIn members declaring to have AI skills (OECD, 2024[29]). AI engineers and AI consultants ranked first and second, respectively, among the fastest growing jobs in the United States between 2022 and 2024 (LinkedIn News, 2025[30]).
Between-country AI skills migration, 2019 and 2024
Source: OECD (2025[31]), “Live data: AI Jobs and Skills”, OECD AI Policy Observatory, https://oecd.ai.
Using Lightcast data for ten OECD Member countries, Green (2024[32]) reveals that, on average, one-third of job vacancies in 2021‑2022 were in occupations with high exposure to AI2 (Figure 5). This finding underscores the growing pervasiveness of AI-related technologies across a broad spectrum of tasks. While this affects traditionally technology-intensive sectors, it also increasingly touches sectors such as finance, healthcare, marketing and administrative services. As shown in the previous section, high AI exposure does not necessarily imply imminent automation. Rather, it reflects the likelihood that AI tools could significantly augment or transform the nature of these jobs.
Job vacancies with high exposure to AI as a percentage of all vacancies, 2021-2022
Note: Vacancies with high exposure to artificial intelligence have an occupational artificial exposure measure one standard deviation greater than the mean according to Felten, Raj and Seamans (2021[33]).
Source: Green (2024[32]), Artificial intelligence and the changing demand for skills in the labour market, https://doi.org/10.1787/88684e36-en.
The most demanded skills in occupations with high AI exposure are management, business processes and complementary skills. Management and business processes contain skills such as project management, budgeting and accounting, administration, clerical tasks and customer support. On average across countries included in the sample, 72% of vacancies in high AI-exposure occupations demand at least one management skill and 67% demand at least one business processes skill. Social, emotional and digital skills are also highly demanded with over half of vacancies in high-exposure occupations demanding at least one skill from these skill groups. High-exposure occupations comprise a third of all vacancies considered in the sample (Green, 2024[32]).
In parallel, growing use of algorithmic management tools in the workplace is increasing the need for certain skills, such as those related to general management and analysis (e.g. ability to use or interpret data) (Milanez, Lemmens and Ruggiu, 2025[34]). Similarly, SMEs indicate that GenAI has made data analysis and interpretation skills more important, along with other skills (OECD, 2025[11]).
As firms increasingly adopt AI technologies, one emerging issue is the widening skills gap, particularly as the demand for high-skilled workers intensifies. AI adoption often brings with it a shift in task composition. It automates routine and repetitive activities, while amplifying the need for advanced analytical, technical and cognitive skills. This dynamic creates a structural challenge in the labour market: individuals with lower levels of education and limited digital competencies face significantly higher risks of job displacement or stagnant career progression (Lane, 2024[16]). Many of these workers are concentrated in sectors such as manufacturing, retail and administrative support, where automation is expected to have the greatest impact.
For Acemoglu and Restrepo (2018[14]), the widening skills gap contributes to a growing polarisation of the labour market. In this environment, middle- and low-skill jobs progressively shrink, while high-skill, high-wage roles (often in technology, finance or professional services) expand. This results in a hollowing-out effect, leaving a gap between high-paying knowledge jobs and low-wage service work, with fewer opportunities for mobility in between. Without targeted interventions, this polarisation may deepen social and economic divides, especially affecting younger workers; older adults with outdated skillsets; and historically disadvantaged groups.
Persistent skills gaps in AI contribute both to overall disparities in adoption and productivity, and to differences across sexes and regions. As in the case of broader ICT specialist occupations, women remain underrepresented in AI-related research, accounting for only 25% of AI researchers. Similarly, women make up just 11% of sole-authored publications (Caira, Russo and Aranda, 2023[35]), and employment in AI-intensive sectors (OECD, 2023[2]). At the same time, AI development and adoption are geographically concentrated in urban and technology hubs, where access to infrastructure, investment and skilled talent is stronger. In contrast, rural areas often lack these resources, making them less likely to benefit from the productivity and innovation gains of AI (OECD, 2024[17]).
← 1. LinkedIn categorises AI skills into two mutually exclusive groups: “AI engineering” and “AI literacy” skills. AI talent refers to “AI engineering” skills. LinkedIn members are considered having an AI talent if they have explicitly added at least two AI engineering skills to their profile and/or they are (or they have been) employed in an AI job.
← 2. The AI exposure measure developed by (Felten, Raj and Seamans, 2021[33]) assesses the extent to which an occupation is likely to be affected by artificial intelligence based on the tasks it involves, rather than on whether AI is currently used in that occupation. The measure links information on AI capabilities with detailed descriptions of occupational tasks from the O*NET database. Occupations receive a higher exposure score when a larger share of their tasks overlaps with capabilities that current AI systems can perform or support, such as processing text, analysing information, or recognising patterns. Importantly, a high AI exposure score does not imply that workers in these occupations are being replaced by AI or that they require specialised AI skills. Instead, it indicates that AI has greater potential to transform or augment the way work is performed within those occupations.