Ursula von der Leyen calls for a single approval procedure for autonomous vehicles so that companies can use approved technologies throughout the European Union, without additional barriers. The Commission President presented the objective at the WinWin summit in Brussels on November 3, as part of her priorities for developing artificial intelligence in industry. The initiative concerns expanding the use of these technologies on the European market; the speech does not adopt a new legal mechanism and does not specify when it would become operational.
In brief Von der Leyen calls for a single approval procedure, followed by the possibility of using the technology throughout the EU. The objective is formulated politically, without a new authorization established through the speech. The EU already has a common system of technical type approval and rules for fully automated driving systems. These do not eliminate all national rules concerning traffic and the operation of driverless services. The Commission’s 2025 automotive plan provided for harmonizing testing on public roads and facilitating cross-border use. The November statement continues this direction without detailing its legal instrument. The cross-border testing initiative began in June with 18 EU countries. After the accessions announced in September, the Commission indicated 22 participating countries, including Norway. A common authorization should respect the conditions under which the technology was verified: road type, area, speed, and operating environment. Approval of a system does not grant it the right to travel without restrictions on any road.
In the speech published by the Commission, von der Leyen links the simplification of approvals to the ability of European manufacturers to develop for a continental market. The idea is that a company should not have to overcome additional administrative barriers each time it wants to use an approved technology in another country. She presents autonomous vehicles as an example of the application of AI in the real economy and argues that Europe should design and produce them, not merely use them.
The novelty is this political support for a single procedure, not the discovery of a legislative gap. The EU already recognizes European type approvals granted under the common framework by the competent authority of a Member State. Type approval determines whether a type of vehicle or a system meets the applicable technical requirements; using a driverless service on a particular route raises other issues as well.
The European type-approval framework allows compliant vehicles to access the market under the conditions laid down by legislation. The manufacturer does not have to repeat the entire European type-approval procedure in each country for the same type of vehicle. This principle should not be confused with granting general permission for every form of autonomous driving and every transport service.
Specific technical requirements exist for fully automated vehicles, introduced in 2022 and subsequently updated. They cover verification of the driving system, response to emergency situations, and the conditions under which it can operate safely. The framework was also developed for applications such as automated parking. Such type-approval amendments are distinct from harmonizing all authorizations for use on public roads.
European rules preserve Member States’ right to regulate traffic and the safety of operating fully automated vehicles in traffic, including in local transport services. Therefore, the vehicle’s technical compliance and permission to operate a particular service are two different levels. For a company expanding across borders, the latter may involve different procedures and conditions.
A vehicle may be verified for transport on a predetermined route, within a defined area, or in a parking facility, without the results demonstrating that it can operate on any European road. In regulation, the system’s operational domain describes the conditions for which it was designed, including the road environment and certain limits concerning speed, traffic, or weather. A common procedure should preserve these limits, even if it reduces the repetition of equivalent checks.
The direction had also been established in the Action Plan for the automotive industry, presented by the Commission in March 2025. The document identified the fragmentation of national rules as an obstacle to the testing and commercial use of autonomous vehicles. It provided for large cross-border testing areas, harmonized procedures for tests on public roads, and rules to facilitate the use of technologies throughout the EU.
The plan included both the development of technical standards and cooperation with countries, local authorities, road operators, and companies. These are complementary directions: a vehicle must meet safety requirements, while the service in which it is used must be able to operate under the traffic conditions for which it is authorized. Announcing these objectives is not equivalent to adopting all the measures provided for.
Cooperation advanced through the declaration of intent for cross-border testing, initially signed by 18 Member States in June. The Commission then announced that Spain had joined the initiative, followed by Portugal, Slovakia, and Norway. The September update thus indicates 22 participating countries. The number does not represent 22 EU Member States, because Norway is not a member of the Union.
The initiative targets applications in public transport, freight transport, and logistics. Participants are working on common principles and coordinated authorization procedures, alongside concrete testing activities and preparations for commercial use. The cooperation declaration alone does not create an authorization valid in all participating countries or throughout the EU.
For manufacturers and operators, the stake of harmonization is reducing uncertainty: which checks are recognized, what conditions must be met in another country, and what obligations arise when the service crosses a border. Reducing duplication could facilitate investment and the expansion of services, but the speech does not provide an estimate of the administrative savings or the exact conditions under which they would arise.
The scope of a possible single procedure remains to be determined. Testing on public roads, urban passenger transport, and freight transport between logistics centers have different operational needs. The speech does not specify which authority would issue the approval, which national permits it would replace, or in what situations additional local conditions could be imposed.
For companies, a market with more coherent rules could make it easier to move from pilot projects to commercial services. For passengers and other road users, the technology’s usage limits, oversight, and the possibility of intervention when problems arise remain essential. The objective presented by von der Leyen is to reduce administrative fragmentation; the legal and safety conditions of a single authorization still need to be specified.
Context: Comisia Europeană și lideri internaționali împing testarea și verificarea modelelor AI, pe fondul îngrijorărilor privind siguranța și controlulÎnaintea discuțiilor UE despre extinderea utilizării tehnologiilor AI, Comisia și-a concentrat atenția asupra riscurilor și a necesității unor evaluări independente pentru modelele foarte avansate. Președinta Ursula von der Leyen a avertizat asupra „riscurilor enorme” și a cerut dezvoltarea capacităților europene, împreună cu cooperare cu parteneri precum Canada și Regatul Unit pentru evaluarea, verificarea și securitatea AI.
Separat, lideri și reprezentanți din 20 de țări, alături de Ursula von der Leyen, au cerut teste obligatorii înainte de lansarea modelelor AI avansate și acces suficient pentru evaluatorii independenți. Apelul propune inclusiv explorarea unei instituții internaționale pentru standarde și verificări, iar inițiativa a fost lansată în perspectiva discuțiilor din cadrul Adunării Generale a ONU.
În paralel, experți CEPS au recomandat ca politica europeană privind riscurile să nu se limiteze doar la scenarii „catastrofale”, ci să combine analiza prejudiciilor deja observabile (dezinformare, discriminare algoritmică, drepturi fundamentale) cu scenarii rare, dar cu impact potențial foarte mare, folosind dovezi, probabilitate și amploarea consecințelor.