EU AI Adoption Doubles Across Strategic Industries, but Skills Gaps Hold Businesses Back

Artificial intelligence adoption has more than doubled across the European Union’s strategic industries since 2021, but a shortage of specialised skills and uncertainty over regulation continue to limit how deeply businesses integrate the technology into their operations.

A new analysis published by the European Commission’s Joint Research Centre on 14 October found that the share of companies using at least one artificial intelligence technology across the EU’s strategic sectors increased from 7% in 2021 to 18% in 2025. Across the wider group of business sectors covered by Eurostat’s enterprise survey, adoption reached 20% in 2025, up from 8% in 2021.

The findings highlight an important distinction for European businesses: adopting AI tools is becoming more common, but using them to transform core business processes remains a separate challenge. The difference could influence productivity, competitiveness and employment as companies across Europe adjust to increasingly accessible AI systems.

AI Adoption Is Growing, but Progress Varies by Industry

The European Commission’s analysis covers ten strategic sectors associated with the EU’s Apply AI strategy. These include media, energy, electronics, pharmaceuticals, defence, climate-related activities, manufacturing, agri-food, tourism and mobility.

Although every sector recorded growth between 2021 and 2025, the pace of adoption differed considerably.

Media and creative industries recorded the highest adoption rate in the strategic-sector analysis, with 38% of firms using AI in 2025. Energy followed at 34%, while the electronics sector reached 30%.

Pharmaceuticals recorded 23%, while defence reached 20%. Climate-related activities and manufacturing each recorded 16%.

At the lower end, agri-food and tourism both reached 13%, while mobility recorded 12%.

These figures show that AI adoption is not progressing uniformly across the European economy. Industries that can apply general-purpose AI to information processing, communication, research and other digital tasks may find it easier to introduce available tools than businesses whose operations depend heavily on physical infrastructure, specialised equipment or complex supply chains.

However, lower adoption does not necessarily mean that a sector has little potential to benefit from AI. It may reflect the cost of implementation, the nature of the work involved or the additional investment required to integrate the technology safely and effectively.

Why AI Adoption Is Accelerating

More accessible general-purpose AI tools

The Joint Research Centre identifies the wider availability and lower cost of large language models as an important driver of adoption.

These systems can analyse and generate text, assist with coding, summarise documents and support the creation of images, audio and video. Their accessibility allows businesses to experiment with AI without necessarily developing a specialised system from the ground up.

For example, a media company might use AI to assist with research organisation, transcription or content workflows. An energy business might explore AI for information analysis or operational support, while an electronics manufacturer could investigate applications in research and development.

These are possible uses rather than evidence that every company in those sectors has implemented them.

The broader point is that businesses can now explore more AI applications through ready-to-use services, although the suitability of each application depends on its purpose, risks and operating environment.

Adoption does not automatically mean transformation

The European Commission’s analysis makes an important distinction between using AI and integrating it into production or other core business processes.

A company might use an AI assistant to draft routine communications without changing how its principal services are delivered. Another company might redesign parts of its workflow, introduce new quality-control procedures and train employees to incorporate AI into daily operations.

Both companies may count as AI users, but the depth of integration is different.

The Joint Research Centre found a positive relationship between organisational changes and AI use in production. Businesses that reorganise their operations to take advantage of the technology are more likely to integrate it into production processes.

This suggests that buying access to an AI tool is only one part of implementation. Businesses also need appropriate workflows, employee training, oversight and clear objectives.

Skills Shortages and Regulatory Uncertainty Remain Major Barriers

The European Commission’s analysis identifies a lack of relevant expertise as a leading obstacle to AI adoption across most of the strategic sectors examined.

Companies may struggle to recruit specialists who understand machine learning, data management, cybersecurity and the integration of AI into existing systems. Smaller businesses may face additional constraints because they have fewer resources to devote to experimentation, staff development and technical implementation.

Regulatory uncertainty is another reported concern, particularly where companies must consider data protection, privacy and the appropriate handling of sensitive information.

For businesses, the challenge is to understand which rules apply to their specific uses of AI and how to establish reliable safeguards. The risks associated with processing confidential customer information, for example, differ from those involved in using AI to generate non-sensitive marketing material.

The appropriate response is not simply to avoid AI. It is to assess individual applications, understand the applicable requirements and introduce controls proportionate to the risks.

Training also matters. Employees need to understand both the capabilities and limitations of AI systems, including the possibility of inaccurate outputs, biased results and inappropriate disclosure of confidential information.

What the Findings Mean for Small Businesses

The figures have particular relevance for small and medium-sized enterprises, which may have fewer financial and technical resources than large corporations.

Eurostat’s 2026 digitalisation publication reports that 20% of EU businesses used AI technologies in 2025. The share was approximately 55% among large enterprises, compared with 19% among small and medium-sized enterprises.

This difference highlights the uneven distribution of adoption across businesses of different sizes.

Smaller companies do not necessarily need to begin with expensive custom-built systems. Depending on their needs, they may be able to assess accessible tools for administrative tasks, customer-service support, document processing or other clearly defined applications.

However, a low initial cost does not remove the need to examine data-handling practices, reliability, contractual conditions and the possibility of human error.

A sensible implementation strategy begins with a specific business problem rather than the technology itself. Companies can identify a repetitive task, assess whether AI is suitable, measure the results and expand only when the benefits justify the additional cost and risk.

This approach can help businesses avoid spending money on tools that produce little practical value.

Potential Effects on Employment and Productivity

The growth in AI adoption raises questions about how work will change across Europe.

AI systems can assist with routine information processing and some repetitive tasks, potentially allowing employees to devote more time to activities requiring judgement, communication, specialist knowledge or direct human interaction.

At the same time, particular tasks may be automated or reorganised, changing the skills employers seek and the responsibilities attached to some jobs.

The adoption figures alone do not establish how many jobs will be created, eliminated or transformed. Those outcomes depend on the applications businesses choose, the pace of implementation, worker training and wider economic conditions.

For employees, the practical priority is to develop skills that complement the technology. These may include digital literacy, critical evaluation of AI-generated information, data protection awareness and the ability to apply professional judgement to automated outputs.

For employers, successful adoption requires more than acquiring software. It also involves communicating changes clearly, training staff and establishing accountability for decisions supported by AI.

How Europe’s AI Policy Could Shape Business Adoption

The European Union is seeking to strengthen its technological competitiveness through initiatives that encourage AI development and adoption across important economic sectors.

Its Apply AI strategy is intended to support wider integration of AI into European industries. The Joint Research Centre’s sector-by-sector analysis provides evidence that can help policymakers identify where adoption is progressing and where additional support may be needed.

A single approach may not work equally well across all industries. Media businesses, energy providers, agricultural companies and manufacturers have different technical requirements, operational risks and investment constraints.

For instance, an industry that primarily uses AI for content generation may require different support from one seeking to integrate AI into complex industrial processes.

The analysis therefore points towards sector-specific measures, skills development and organisational change rather than assuming that access to general-purpose AI tools will be sufficient.

For businesses operating across the EU, regulatory compliance and data protection will remain important considerations alongside cost, productivity and the suitability of particular systems.

What Businesses Should Watch Next

The latest findings provide a useful starting point for understanding the direction of AI adoption in Europe, but the next stage will depend on whether companies move beyond experimentation.

Three developments deserve attention.

First, the integration gap: Businesses and policymakers will need to assess whether AI is becoming embedded in core operations rather than being used only for occasional tasks.

Second, workforce skills: Training and access to relevant expertise will influence how effectively companies can deploy AI and manage its risks.

Third, differences between industries and business sizes: Future progress will depend partly on whether smaller firms and sectors with lower adoption can overcome barriers related to cost, skills and technical requirements.

The findings published in October 2026 show that AI is becoming more widespread across Europe’s strategic industries, but adoption rates alone do not measure the full economic benefit.

For European businesses, the central practical lesson is to focus on useful applications, prepare employees and integrate the technology carefully into existing operations. The companies best positioned to benefit will need to demonstrate not simply that they use AI, but that it helps them work more effectively and responsibly.

Sources

European Commission Joint Research Centre — AI Adoption More Than Doubles in Almost Every EU Strategic Sector, 14 October 2026.

European Commission Joint Research Centre — Measuring AI Adoption in the European Union: Methodology and Results for Strategic Sectors, 2026.

Eurostat — Digitalisation in Europe: 2026 Edition.

Eurostat — The Use of Artificial Intelligence Technologies in the European Union: Key Results, 2026 Edition.