Self‑driving cars in the EU Connected and automated vehicles

Introduction

The European Union seeks to transform its automotive industry to build a cleaner, safer and more efficient transport system, and to make the EU a world leader in connected and automated mobility. Road transport accounts for nearly three quarters of EU overall transport greenhouse‑gas emissions, and is a major source of air pollution in cities. Connected and automated vehicles (CAVs) could help reduce fuel consumption and transport emissions by enabling smarter traffic management, among other things. In 2022, about 20 600 people were killed in road crashes in the EU. CAVs could help reduce accidents caused by human error, which is involved in 95 % of road accidents. Such vehicles could also remove barriers to personal transport and ease mobility for all, supporting inclusive mobility for elderly or disabled people.

As far as global competitiveness is concerned, the CAV market is expected to grow exponentially, with the potential to create jobs and economic opportunities. A 2026 study by the Commission Joint Research Centre finds that the economy‑wide effects of CAVs would depend on whether they are deployed mainly as private vehicles or as part of shared mobility services. The study finds that deploying CAVs through shared mobility services could produce more positive outcomes, including fewer accidents, less congestion and increased energy efficiency.

Intelligent vehicles of the future? Connected and automated

Connectivity and automation are two concepts that are often confused but reinforce each other. For instance, a vehicle's ability to receive and transmit data (connectivity) can increase autonomous capabilities (automation) and vice versa.

Connectivity refers to the ways in which a vehicle communicates with its surrounding environment, including infrastructure and other vehicles. These communication links are collectively referred to as 'vehicle‑to‑everything' (V2X, see Figure 1).

Communication in connected cars is divided in two main groups:

Data source: Graphic by Samy Chahri, EPRS, 2026.

Automation refers to the level of driving automation reached by a vehicle in the absence of human supervision through on‑board sensors and systems able to operate with little or no human input.While the terms 'automated' and 'autonomous' vehicles are sometimes used interchangeably, they describe different levels of automation. Automated vehicles assist or perform specific driving tasks in the presence of human drivers, while autonomous (or fully automated) vehicles operate without human involvement. When referring to the different levels of automation, most organisations follow the classification developed by the International Society of Automotive Engineers (SAE). The levels of automation identify how the aspects of the dynamic driving task are divided between the human and the vehicle's systems in different driving modes (see Figure 2).

Data source: Graphic adapted from SAE Standard J3016, SAE International 2014, by Samy Chahri, EPRS, 2026 (see related EPRS briefing from 2016).

At level 0 (e.g.the vintage car), the driver must constantly supervise the road, while the fully autonomous car of the future (level 5) can perform all driving functions under any condition without driver input.

At present, CAVs with SAE levels 4 and 5, corresponding to highly and fully autonomous vehicles, are not yet commercially available in the EU, as their introduction requires further technological development and regulatory updates. Nonetheless, vehicles with driver‑assistance systems (levels 1 and 2) are already available on the EU market (e.g. lane‑keeping assistance). Level 3 automated vehicles for low‑speed driving on motorways became commercially available in the EU in 2021.

The rise of artificial intelligence (AI) is playing a key role in helping CAVs to interpret vast streams of sensor and connectivity data and turn them into safe driving decisions. Industry has broadly developed two main approaches to design AI systems for such vehicles:

EU policy objectives

A key EU ambition is to reach large‑scale CAV deployment by 2030. The European Commission's sustainable and smart mobility strategy (SSMS) presents a set of interconnected objectives, primarily centred on safety, competitiveness, sustainability and technological leadership.

Road safety is a key early driver of automation. Automation can contribute to the EU's overarching goal of reaching zero fatalities in road transport by 2050, known as Vision Zero 2050. The vision sets interim targets of reducing both road deaths and serious injuries by 50 % by 2030. While its impact should not be overestimated, automation could reduce accidents caused by human error and increase safety outcomes. A study estimates that by 2040, around 22 % of collisions could be prevented through a robust CAV framework. In its 2021 resolution on the EU road safety policy framework, the European Parliament explicitly called on the Commission to propose 'a new harmonised regulatory framework' for CAVs. The Commission acknowledged the request, but a fully harmonised framework for CAVs is not yet in place.

Beyond safety, industrial competitiveness is a strategic objective in the EU's efforts to position itself as a global leader in connected, cooperative and automated mobility (CCAM). The Commission's 2018 communication 'On the road to automated mobility' outlines the objective, which was also supported by Parliament's 2018 resolution on cooperative intelligent transport systems and its 2019 resolution on autonomous driving. The automotive sector remains a cornerstone of the EU economy. It directly employs more than three million people and accounts for around one third of private research and development investment in the EU. The EU has therefore backed CCAM development through research and innovation funding, including the Horizon Europe CCAM Partnership, which has planned investment of €500 million. The Commission's March 2025 automotive industrial action plan identifies CAVs as a means of strengthening competitiveness and calls for harmonised EU rules, cross‑border test beds, industrial alliances, secure data access, cybersecurity standards, and investment in chips, batteries and digital infrastructure. In June 2026, 18 Member States signed a joint declaration of intent on large‑scale cross‑border CAV test beds, seeking common approval principles, coordinated permitting and deployment, and greater certainty for innovators and investors.

From a sustainability perspective, connectivity and automation are expected to optimise traffic flows, making transport more efficient. As set out in the 2030 climate target plan, increasing the share of connected, automated and multimodal mobility can lower pollution, especially in cities.

Vehicle automation has also been incorporated into broader strategies to secure technological leadership in digital and AI policy. The Commission's 2025 apply AI strategy identifies mobility as a key sectoral flagship. It has also launched the Autonomous Drive Ambition Cities initiative to accelerate the deployment of autonomous mobility technologies and services (e.g. robotaxis, autonomous shuttles and advanced self‑driving cars) in EU cities by 2030. At the same time, the Commission has emphasised the importance of establishing the European Common Mobility Data Space to facilitate secure data pooling and sharing across transport systems, including reducing dependence on non‑European data platforms and ensuring access to real‑time traffic information.

EU legislative framework

Challenges related to CAV deployment

Barriers to large‑scale deployment remain, affecting the competitiveness of the CAV sector.

The EU is relatively strong in research and development (R&D). However, several moral questions and ethical dilemmas remain to be addressed, such as algorithmic bias in pedestrian detection, surveillance through mobility data and the democratic legitimacy of companies making moral decisions about CAV behaviour without public oversight.

Following R&D, demonstration and piloting are key. Level 3 and 4 pilot projects (such as the automated shuttle bus in Leuven) provide learning opportunities. The Commission further identifies several level 3 and 4 use cases, including motorway chauffeur functions, urban shuttles, automated valet parking and freight applications that can help advance to higher levels of automation. The Commission has announced plans to establish a minimum of three large‑scale cross‑border test beds, regulatory sandboxes and European automated driving corridors to improve the conditions for road testing. In June 2026, the Autonomous Drive Ambitious Cities (ADACities) initiative was launched. Participating cities are expected to become real‑world leaders in the deployment of CAVs with target fleets of 100 or more CAVs by 2030.

The EU is strong in the field of semi‑automated vehicles (levels 1 and 2), but hurdles to scaling up production of higher‑level CAVs persist. Mario Draghi's 2024 report notes that the EU automotive industry is weakened by high production costs, lagging technological capabilities and supply chain dependencies. Chinese and US producers dominate the CAV market. R&D capacity does not automatically translate into the capacity to scale up, due to several main gaps, as listed below.

Re‑thinking transport and mobility

Although the full effects of high‑level vehicle automation on EU society, the economy and traffic flows remain uncertain, the technology will likely transform transport systems. Expected benefits include reduced delays and congestion, better accessibility, lower travel costs and new revenue opportunities for ride‑sharing operators. Effects depend heavily on CAV market share and ownership structure. Several factors are expected to increase total road transport kilometres, potentially offsetting some of the benefits, as described below.

The Commission highlights the potential for new mobility services based on CAVs and for the spread of cooperative intelligent transport services. In designated areas such as highways, ports and shuttle routes, highly automated vehicles are already operating.

CAVs can serve both passenger and freight transport. The European Automobile Manufacturers Association (ACEA) identifies six use cases: for passenger transport, automated valet parking, (shared) passenger services and private vehicles on highways, and for freight transport, yard operations, last‑mile delivery, and hub‑to‑hub freight transport.

Policy views

Regarding the processing of personal data, the European Data Protection Board (EDPB) published guidelines on personal data processing in the CAV context in March 2021. Considering the vast volume of personal data produced by CAVs, the EDPB recommends that CAVs follow a 'privacy‑by‑design' approach. Manufacturers and service providers should collect only the data needed for a specific purpose, use privacy‑friendly settings by default, and give drivers and passengers clear information and practical control over their data. Whenever possible, data should be processed inside the vehicle rather than transferred to external servers. Where sharing is necessary, data should be anonymised/pseudonymised and subject to appropriate security safeguards. Special care is needed for location, biometric and data revealing traffic offences, as these may expose sensitive information about a person's habits or behaviour.

The EDPB also considers CAVs to be 'terminal equipment' under the e‑PD. Therefore, in principle, prior user consent (Article 5(3) of the ePD) is required for accessing data stored in the vehicle.⁵ The GDPR – with all the other grounds for processing personal data (Article 6 GDPR) – governs any subsequent processing.

ENISA published a set of AI recommendations addressing cybersecurity challenges posed by CAVs, ranging from the systematic security validation of AI models and data to incident handling and vulnerability discovery. These recommendations include: (i) adopting resilience mechanisms that prepare alternative plans and incident response activities in case of incidents; (ii) establishing a proper AI security policy across the supply chain, including third‑party providers; (iii) promoting the use of standardised components and homogeneous AI solutions in the automotive context; (iv) establishing a culture of learning from AI security incidents and improving information sharing among stakeholders.

Part of academia suggests introducing strict liability for a limited class of high‑risk systems (e.g. CAVs) as necessary to complete the architecture envisaged when the AI Act, the Product Liability Directive and the then withdrawn AI Liability Directive were conceived as an integrated package. The Commission's expert group on liability and new technologies suggested applying strict liability to high‑risk emerging digital technologies that are operated in public environments and may cause significant harm, including autonomous vehicles and drones. As an example, the expert group recommended that the legislator 'could decide that for CAVs with a level of automation of 4 or 5, it is the provider running the system and who enters the AV in the national registry who is liable. This provider would therefore also take out insurance and could pass on the premiums through the fees paid for its services.'

On research and development, STOA recommends combining a limited number of European Centres of Excellence with a distributed network of research institutions, balancing concentrated expertise with broad participation. To address the lack of coordination and harmonisation in the sector, it suggests creating a new European administrative agency for road vehicles to harmonise operational design domain (ODD) criteria, standardise safety cases, coordinate incident/accident data and enable mutual recognition of national operational approvals. Clearer safety standards and communication can be beneficial for public trust and user acceptance.

To address the impact of automation on the transport workforce, Commission recommendation (EU) 2024/236 invites employers, workers' organisations, and national authorities to pursue a transition along five lines: raising awareness of automation's impact, upskilling and re‑skilling the workforce, improving working conditions, managing change proactively and ensuring adequate funding. It places emphasis on social dialogue between employers and workers to shape how automation is introduced.

High levels of shared mobility/CAVs have a greater potential to increase benefits of the new technology: In an ideal scenario, they can reduce congestion and human‑induced traffic collisions by up to 70 %. In its analysis, the French national strategic foresight advisory body encourages a systemic approach to passenger transport mobility. The approach combines express bus and carpooling lines with other services to create a system that is also competitive outside dense urban areas, and it supports local trip sharing initiatives (reserved lanes for sharing services) and favourable parking policies.