Driverless and autonomous vehicles (“AVs”) are moving from test tracks and pilot schemes to commercial deployment around the world. In the U.S. and China, robotaxi fleets and autonomous trucking services are proliferating. Similar developments are happening in the EU and UK. In London, the first commercial autonomous ride-hailing service launched in early September 2026, with a supervised fleet carrying paying passengers. In Germany, regulators have granted the first nationwide authorisations for Level 4 urban testing and cab-less Level 4 autonomous trucks. At the international level, UNECE has adopted the first global rules for fully driverless vehicles, as discussed in our earlier blog post here.
As AV technology matures, the regulatory and liability frameworks that will govern it are starting to emerge. This article provides an overview of the key developments that will affect autonomous vehicle manufacturers, and technology companies and mobility-service providers operating in this sector.
Current State of AV Deployment
In the EU, 18 Member States signed a Joint Declaration of Intent in June 2026 on large-scale cross-border testbeds for autonomous vehicles, under the European Automotive Action Plan. In Germany, the Federal Motor Transport Authority (KBA) granted nationwide authorisation in July 2026 for Level 4 automated-driving testing on urban roads — the first such nationwide urban approval, allowing testing across German cities without a separate authorisation process required in each location.
In the UK, the Automated Vehicles (Permits for Automated Passenger Services) Regulations 2026 came into force on 15 May 2026, creating a clear framework and dedicated portal for companies to obtain permits to offer automated passenger services. The portal is now open. As of late September 2026, no operator has yet completed the relevant vehicle-registration stage or applied for a driverless permit – we continue to monitor this space. Separately, the UK Department for Transport’s consultation on a draft Statement of Safety Principles for automated vehicles, which will set the safety-related benchmarks for automated vehicles, closed on 9 September 2026.
London’s first commercial autonomous ride-hailing service launched in early September 2026. This currently operates with a trained, Transport for London (TfL)-licensed driver on board to monitor trips and intervene if necessary. These services run under the ordinary private-hire licensing framework rather than the new driverless permit scheme.
Implementing the UNECE Framework for Automated Driving Systems
In June 2026, UNECE’s World Forum for Harmonization of Vehicle Regulations (“WP.29”) adopted the first international regulatory framework for fully driverless Automated Driving Systems (“ADS”), comprising a new UN Regulation (“UNR”), and a companion Global Technical Regulation (“GTR”). For more information, see our earlier blog post here.
Following adoption of the UNR on ADS, the European Commission is preparing a delegated regulation to amend the annexes to the General Safety Regulation (EU) 2019/2144, which is likely to incorporate the new UN ADS Regulation. Commission adoption is currently planned for the fourth quarter of 2026.
The UK’s plans for adopting the UNR on ADS is less clear at this stage. The UK Department for Transport’s August 2026 consultation on GB type approval (which closed on 29 September) does not propose incorporating the UN ADS Regulation into the domestic framework.
Liability Framework
As AVs move from test tracks to public roads, the question of who bears liability when something goes wrong remains top of mind for legislators, regulators and industry stakeholders. The EU’s recent reform of its product liability framework will directly affect companies supplying autonomous-driving software and AI, and similar changes may be to come in the UK.
The EU is already ahead of the UK in this area, having adopted a major reform of its product liability framework through the revised Product Liability Directive (“PLD”). Although already adopted, the revised PLD applies only to products placed on the market or put into service from 9 December 2026. The revised PLD modernises EU product liability law by expressly bringing software within scope, whether supplied on a standalone basis or integrated into a product. This is particularly significant for autonomous vehicles, where liability exposure may extend beyond the physical vehicle to autonomous-driving software and AI-enabled driving functions. In addition, certain related digital services may themselves fall within scope, including, for example, services providing the continuous supply of traffic or navigation data. The PLD must be transposed into national law by 9 December 2026. Transposition is underway and divergences are already emerging (creating uncertainty and scope for satellite litigation). The Commission has also been engaging with Member States on implementation issues as the date of application approaches. (For more information on the PLD, see our client alert here, and our blog post here on PLD transposition.)
Meanwhile, a cross-party group of senior MEPs has reportedly recently urged the EU Commission to consider new specific AI liability legislation to address perceived gaps in the framework for harms caused by advanced AI systems, following the formal withdrawal in 2025 of the proposed AI Liability Directive.
In the UK, the Jurisdiction Taskforce (“UKJT”) recently analyzed the current provision for allocation of responsibility for AI-related harms, concluding that existing English law principles (e.g., contract, negligence and, to a lesser extent, statutory strict liability) are generally capable of allocating responsibility without a new bespoke regime. The statement includes an AV-specific illustration, suggesting that existing principles of contributory negligence could be applied to reduce liability where an AI system recommended a driving route across soft tidal sands and a user crossed the beach knowing the hazard. The statement is non-binding but likely to be of persuasive value (the UKJT’s membership includes respected academics and members of the judiciary).
However, the UK Law Commission is actively reviewing the current product liability regime with a view to making recommendations for reform, and has AI clearly in its sights. It considers the current regime to be no longer fit for purpose, and is considering options to modernise the law, including to account for emerging technologies, including AI. The review is considering whether to introduce similar claimant-friendly reforms to those in the new PLD. Its proposals for reform are expected to be published for consultation before the end of 2026.
Next Steps
The remainder of 2026 appears likely to be a defining period for AV regulation in the EU and UK. In the EU, the Commission’s delegated regulation incorporating the UNECE driverless framework into the General Safety Regulation is expected in Q4, while the deadline for Member State transposition of the revised Product Liability Directive is 9 December.
In the UK, the Department for Transport has promised a further consultation in the second half of 2026 on the Automated Vehicles Act secondary legislation, covering authorisation, operator licensing, safety principles and data access. The Law Commission’s product liability reform proposals are expected before year-end. On the deployment side, permits for the first automated passenger-services and paid journeys are yet to be granted, but may not be that far off.
Taken together, these developments indicate that AV regulation in the EU and UK is entering a more operational phase. Businesses across the AV supply chain should be engaging now to closely monitor the rules still being written and to prepare for the compliance obligations that are already or will shortly be in effect.