What the new UN draft means for OEMs and tech players

Ten years ago, automated driving was expected to be common on public roads by 2020. The technology advanced rapidly, but technological maturity alone proved insufficient. Without an internationally aligned regulatory framework, autonomous driving remained confined to pilot projects.

At the end of January, the Working Party on Automated, Autonomous and Connected Vehicles (GRVA) of the United Nations Economic Commission for Europe (UNECE) adopted a draft global regulation on Automated Driving Systems (ADS). After a decade of technological development, regulatory evolution and two years of intensive international consultation, the draft is scheduled for submission to the World Forum for Harmonization of Vehicle Regulations (WP.29) in June 2026. If adopted, the regulation will enter into force immediately.

For the first time, a harmonised global legal framework would enable the deployment of driverless autonomous vehicles on public roads.

What to expect at the Mobility + AI Conference 2026

On 14 and 15 April 2026, the Mobility + AI Conference takes place in Munich. Across eight core themes, the event explores the intersection of technology, safety and regulation – from AI-driven automated driving and robotaxis to autonomous logistics, validation, cybersecurity and SDV architectures.

Benjamin Müller

Harmonisation as an industrial policy lever

The implications extend far beyond technical details. Harmonised vehicle regulations aim to prevent fragmented national approaches, enable economies of scale and facilitate both innovation and market access.

International reactions underline the strategic relevance. The US National Highway Traffic Safety Administration (NHTSA) has launched a public consultation on the draft. China has announced that it intends to align its national standard with the global framework. Japan and several European countries have explicitly welcomed the initiative.

For OEMs, the UN draft is emerging as the global reference architecture for autonomous driving. Early regulatory alignment becomes a prerequisite for international scalability.

From driver assistance to full system responsibility

At the core of the regulation is the definition of ADS as systems that perform the entire dynamic driving task within their Operational Design Domain (ODD). This includes perception, decision-making and vehicle control.

The regulatory principle is explicit: within its ODD, an ADS must operate at least as safely as a competent and careful human driver.

This shifts the focus fundamentally. The decisive question is no longer whether the system can technically drive. It is whether the manufacturer can demonstrate that it drives safely. Proof replaces promise.

Safety case becomes central to type approval

The draft embeds the safety case approach as a core instrument. Manufacturers must provide structured evidence of hazard identification, risk assessment and risk mitigation. They must explain how the system transitions to a safe state in case of malfunction and demonstrate that no unreasonable residual risks remain. Safety becomes a documented argumentation architecture.

In addition, a mandatory Safety Management System (SMS) is introduced. Safety must be organised, documented and auditable throughout the entire lifecycle. Processes, competencies and responsibilities become subject to regulatory review. The SMS itself requires assessment and certification.

For OEMs, this means safety is no longer solely a technical discipline. Artificial intelligence development, functional safety, cybersecurity and homologation must be structurally integrated. Organisational silos become regulatory risk factors.

Validation must withstand methodological scrutiny

The global regulation does not prescribe a single test method but establishes a harmonised evaluation logic. Simulation, virtual toolchains, test bench procedures and real-world driving must be systematically combined.

Manufacturers must demonstrate that their test environments are representative, that relevant scenarios are adequately covered and that virtual results are transferable to real-world conditions. For AI-based systems, functional performance alone is insufficient. The validation process itself must be transparent, traceable and auditable.

For software and AI teams, this makes traceability between training data, scenarios, safety goals and system behaviour a regulatory requirement. Simulation becomes a strategic core capability. Black-box approaches without explainable safety arguments face increasing pressure.

Regulation extends into operation: monitoring becomes mandatory

Waymo boosts war chest in robotaxi race

In a fresh funding round, Waymo has secured significant new capital to support its robotaxi ambitions. The Google sister company could further extend its lead in the race with Tesla for autonomous ride-hailing services.

Benjamin Müller

One of the most far-reaching changes concerns the post-approval phase. The introduction of In-Service Monitoring and Reporting requirements (ISMR) makes continuous performance monitoring part of the regulatory framework.

Manufacturers must collect field performance data, analyse safety-relevant events and report regularly to authorities. A Data Storage System for Automated Driving (DSSAD) becomes mandatory, recording safety-related ADS performance data.

Regulation no longer ends at market entry. It accompanies the product throughout its lifecycle. Data architecture, over-the-air update strategies and software governance become homologation-relevant.

Concrete implications for the industry

For OEMs, the new framework represents a strategic turning point. Defining the ODD becomes central to product strategy. Each expansion increases validation effort and regulatory complexity. At the same time, global harmonisation opens opportunities for international scaling.

Suppliers, particularly those providing sensors, computing platforms and software stacks, must deliver regulatory-ready development processes, safety analyses and documentation. Performance alone is no longer sufficient. The role shifts from technology provider to compliance co-architect.

Technology companies and AI specialists must ensure that toolchains and training processes are auditable. Continuous learning in the field becomes subject to regulatory oversight.

According to Daniel Fulger, Global Director AI Automotive, Engineering & Homologation at TÜV Rheinland, safety, innovation and public trust must evolve together, without rising development costs being compounded by equally rising homologation burdens. Ultimately, global competition includes markets that allow self-certification.

Hamburg and Free Now by Lyft launch robotaxi pilot

Hamburg is taking a first step towards integrating autonomous taxis into its urban transport system. Together with Free Now by Lyft, northern Germany’s largest city is launching a pilot project designed to test Level-4 robotaxis under real-world conditions.

Dr. Martin Large