The OECD’s 2025 AI Capability Indicators Technical Report introduces a comprehensive international framework to systematically assess what artificial intelligence systems can and cannot do compared to human abilities. Developed under the AI and Future of Skills project, the initiative establishes ten capability indicators (language, social interaction, problem-solving, creativity, metacognition and critical thinking, learning and memory, vision, manipulation, robotic intelligence, and consciousness) to evaluate AI systems through a five-level scale ranging from basic to human-equivalent and beyond. Each indicator describes the degree to which AI systems demonstrate complex cognitive and physical abilities, from perception and reasoning to social behaviour and creativity.
The framework was designed to provide policy-makers, educators, and researchers with an evidence-based tool for understanding AI’s real-world performance and its implications for employment, skills, and governance. It addresses a critical gap between technical benchmarking and human comparability by integrating expert judgement with empirical AI performance data, drawing on over thirty leading scientists and psychologists worldwide. Results indicate that AI systems show the highest maturity in language processing, computer vision, and data analysis, while lagging in social intelligence, creativity, and autonomous physical interaction. The report also introduces an “AI Catch-Up Index” that links capability indicators to occupational data, providing insights into which human abilities are most susceptible to automation or transformation.
Beyond mapping AI’s technical progress, the indicators aim to inform strategies on workforce adaptation, education reform, and ethical governance. The OECD emphasises the need for continuous updating of the scales to reflect rapid technological advances, including generative AI and agentic systems, and calls for coordinated international collaboration to ensure that AI development remains transparent, human-centred, and globally accountable. Ultimately, the framework serves as a foundation for measuring AI competence, human–AI complementarity, and societal readiness, providing governments with the tools to align innovation policy, labour-market foresight, and digital skills development with the realities of a fast-evolving AI landscape.

The Joint Research Centre’s 2025 report AI Skills Supply and Demand analyses the match between AI-related education and training in Europe and labour market demand using data from Studyportals and online job advertisements. It finds a general alignment between the two, but both remain heavily concentrated in ICT fields, with weak integration of AI skills in areas such as health, agriculture, and social sciences. Most EU master’s programmes and short courses focus on machine learning (32%), generic AI (26%), and AI ethics (10%), while generative AI accounts for less than 2%, indicating a lag in addressing emerging technologies.

Short professional courses, often linked to micro-credentials, are increasingly used for rapid reskilling and upskilling, helping professionals adapt to fast-evolving AI roles. On the demand side, AI-related job postings are dominated by software developers, data analysts, data engineers, and AI/ML engineers, which together make up nearly all AI-specific roles. The study observes that AI ethics appears in education but rarely in job ads, suggesting universities are ahead of industry in anticipating future skill needs.

The report concludes that while Europe’s AI talent pipeline is broadly aligned with demand, future gaps could emerge if AI training remains confined to ICT. It calls for multidisciplinary integration of AI across sectors, modular learning pathways, and ongoing monitoring of generative AI and ethical governance skills to sustain Europe’s competitiveness and technological sovereignty.
 

UNESCO’s 2025 report Media and Information Literacy for All: Closing the Gaps offers the first comprehensive global overview of how countries are integrating media and information literacy (MIL) into education systems and national policy. While 171 of 194 UNESCO Member States reference MIL or related competencies in national frameworks, implementation remains uneven and fragmented. Only 43% of countries have integrated MIL into school curricula, and another 29% focus solely on digital-skills training. Europe and North America lead with 91% integration, while Africa, Asia, and Latin America continue to lag, revealing persistent global inequalities in access to critical information skills.

The study finds that MIL is most frequently taught at the secondary level, often embedded across multiple subjects rather than offered as a standalone course. This cross-curricular approach helps connect MIL to real-world issues but makes progress difficult to track and sustain. Countries that have developed standalone or hybrid MIL policies (only 17 worldwide) show higher success rates in embedding critical thinking, media analysis, and ethical digital engagement in education. By contrast, systems that limit MIL to technical digital literacy risk neglecting the broader competencies needed for navigating misinformation and AI-driven content.

UNESCO concludes that a global policy–practice gap persists, with limited funding and fragmented coordination slowing progress. The report calls for stronger national strategies, sustained investment, and multi-stakeholder partnerships linking governments, educators, civil society, and international bodies. It recommends treating MIL as a core twenty-first-century competency, essential for democracy, inclusion, and resilience in an increasingly complex information environment.