Reshaping Europe’s workforce in light of AI
Recent research suggests roughly a third of global work hours could be automated by the end of the decade. With advances in artificial intelligence (AI), many monotonous and repetitive aspects of today’s work currently done by humans could be completed by computers. “Tasks most amenable to AI and automation are typically routine, repetitive and rule-based,” explains Meltem Ucal, professor of Economics at Kadir Has University. These include activities such as data entry and validation, document classification and processing, standardised reporting, scheduling and coordination, repetitive monitoring or triage tasks, and first-line responses to predictable customer queries. “AI systems can accelerate these processes, reduce errors and free human workers to focus on tasks, creativity and interpersonal skills,” says Ucal. In light of this, in the AI4LABOUR(opens in new window) project, which was funded by the Marie Skłodowska-Curie Actions(opens in new window) programme, Ucal and her team worked to predict emerging occupations and the required training needed to succeed in these jobs. The project used a novel AI model to create an integrated framework that links jobs, tasks, skills and training pathways and created a series of open recommendations.
Analysis of future occupations
Methodologically, AI4LABOUR combined several strands. Using task- and skill-based modelling, the researchers first analysed occupations at the task level rather than only job titles. With machine learning and deep learning approaches, they could then apply classifiers and deep learning techniques to distinguish routine versus non-routine tasks and estimate automation probabilities. This was complemented with survey-based data collection, using scientifically designed surveys to generate empirical evidence from employees across partner organisations. The researchers then validated their model and evaluated its accuracy before mapping model outputs to reskilling or upskilling pathways. Finally, they created a recommendation portal which translates models and datasets into a practical recommender system.
Understanding labour market change at the task level
An important insight from the project was that labour market change is best understood at the task level. “Many occupations are not simply ‘disappearing’, but transforming through task recomposition,” explains Ucal. “Future demand frequently lies at the intersection of domain expertise and digital and AI literacy.” Specific outcomes include: an AI-enabled task–skilling approach to assess occupational exposure to automation; validation of machine learning / deep learning models and accuracy evaluation reports; a gender-focused analysis highlighting fairness considerations; and the development of the web-based recommendation portal. “The AI4LABOUR recommendation portal is designed to be publicly accessible, allowing individuals and stakeholders to visit the website and use its predictive and training recommendation functions,” notes Ucal. “The project has also generated a variety of public outputs, including reports, publications, and openly available research products and a thesis.”
Preparing the European workforce
Rather than reactive adjustment, AI4LABOUR aimed to support anticipatory adaptation to incoming changes. By identifying task profiles more exposed to automation and linking them to training pathways, the work can help individuals plan reskilling, companies design workforce strategies, educators align curricula, and policymakers develop targeted labour market policies. “The overall goal is smoother, fairer transitions and improved employability,” says Ucal. The researchers will continue to refine the models, expand their datasets and explore follow-up collaborations. “Further development and long-term sustainability of the portal concept remain priorities as technologies and skills demands evolve,” adds Ucal.
