Are they legal yet? International automated vehicle regulation

With deployments of automated vehicles (AVs), in particular robotaxis, expanding internationally, governments are seeking to legislate for commercial deployments. 2026 is looking like a significant year for automated vehicle regulation, with progress internationally and across jurisdictions. This article provides an update on recent international developments and discusses some key issues – level of safety, remote operators and Level 3 features.

Draft UNECE Regulations

The United Nations Economic Commission for Europe (UNECE) sets international technical standards for road vehicles. Over recent years, working groups have developed requirements for automated driving systems (ADSs). These have now come together as proposals for a new draft United Nations (UN) “Regulation on uniform provisions concerning the approval of motor vehicles with regard to their Automated Driving Systems” and a draft UN Global Technical Regulation on Automated Driving Systems.

The proposed regulations are very comprehensive. The documents do not set requirements for specific applications or operational design domains. Instead, they provide a framework for demonstrating safety and compliance of ADSs. The proposed regulations cover both passenger and commercial vehicles and “establish obligations on manufacturers to:

• implement and maintain a Safety Management System (SMS) addressing the entire lifecycle of the ADS,
• substantiate a comprehensive safety case and
• demonstrate the credibility of the testing environments and methods used.”

Some key features of the draft documents include:

• a proposed safety standard of a “careful and competent driver”, whilst the ADS must also be “free of unreasonable risk”;
• requirements for manufacturers to develop a safety management system (including a focus on “safety culture”) along with a safety case for the specific ADS;
• the ADS must comply with local traffic laws, detect and respond to priority vehicles and “respond to instructions from road safety agents”;
• requirements for Level 3 and Level 4 features (although they introduce new terminology* for these features);
• requirements for demonstrating virtual, on-track and on-road testing;
• requirements for data recording and in-service monitoring and reporting (“ISMR”), including reporting timeframes and information required for critical, significant and other “occurrences”.

A huge amount of work has gone into these drafts. Whilst there will likely be further debate on the detail (and some consistency issues to work through), these documents are a major step towards bringing AV technology into mainstream international vehicle regulation. Once implemented, they will require significant resourcing from national regulators to manage approvals and reporting.

The documents will be considered at the next meeting of Working Party 29 later this month (23-26 June).

Singapore

In May, Singapore put out a consultation paper seeking feedback on key aspects of AV regulation as they move to broader commercial deployments. The Singapore government is seeking feedback on key roles and responsibilities; liability, compensation and insurance; data and cybersecurity; and the regulation of advanced driver assistance systems (ADAS).

The Singapore consultation raises some important issues and questions, including:

• ensuring there is a single entity in charge of AV technology with these entities potentially to be licensed, including corporate presence and financial capability requirements;
• whether remote operators should be regulated;
• whether Level 3 automated driving systems should be allowed and if so, how they should be regulated;
• statutory presumptions/ burden of proof for liability and compensation mechanisms.

The consultation is open until 30 June 2026. The paper notes the intention to introduce an “AV Act” in 2027.

China

In April this year, China’s Ministry of Industry and Information Technology published national “Safety Requirements for Automated Driving Systems of Intelligent Connected Vehicles” (Chinese language version) for public comment.**

The requirements cover both passenger and freight vehicles and Level 3 and Level 4 ADS features. The document broadly aligns with the developing UNECE standards. This includes requirements for a safety case, a safety management system and a data storage system for automated driving. The requirements propose that “the safety performance of the ADS shall be at least equivalent to that of a competent and attentive human driver.” The ADS also “shall not impose unreasonable safety risks to users or other road users.” The requirements include provisions for remote assistance; training for fall-back users; and a requirement for continuous safety improvement.

The consultation period has finished; however a final document does not yet appear to have been published.

The United States

The United States (US) continues to see a patchwork of regulation at state and city level, but there has been some movement this year towards federal regulation. At the State level, California has recently released new AV regulations to cover heavy vehicles and provide more detailed requirements on enforcement and interactions with emergency services.

The US National Highway Traffic Safety Administration (NHTSA) has published proposed rulemakings proposing minor amendments to vehicle standards. These remove some requirements for manual controls, windscreen wipers and defrosting systems. The NHTSA has also sought public comment on a request for exemption for Zoox’s automated passenger vehicle, including the use of rear-facing seats. The NHTSA has also indicated that it will be releasing new guidance this year on automated vehicle safety.

Congress has introduced a new bill (the Safely Ensuring Lives Future Deployment and Research In Vehicle Evolution Act of 2026’’ or ‘‘SELF DRIVE Act of 2026’’). The Bill covers “ADS-equipped” vehicles with Levels 3, 4 and 5 features. It includes the term “ADS-dedicated vehicle” for a vehicle designed to be operated exclusively at Level 4 or Level 5.

The Bill would require manufacturers to develop a safety case for an ADS along with a cybersecurity plan. It would empower the NHTSA to develop safety standards for ADS-equipped vehicles and detailed requirements for the safety case. It requires that the safety case show that the ADS-equipped vehicle “will not present an unreasonable risk of accidents, death, or injury”. Also included in the Bill are:

• a requirement that the ADS must “comply with any applicable State or local traffic law and any law relevant to the performance of the DDT.”
• exemptions from requirements for manual controls.

Finally, the Bill proposes to establish a “National Automated Vehicle Safety Data Repository” for AV crash information. The Bill proclaims Federal pre-emption, in particular over any state laws that would ban AVs.

There have been several unsuccessful attempts at Federal legislation in the US, so it’s unclear whether this Bill will proceed, but it has been welcomed by industry.

The United Kingdom (UK)

The UK passed their Automated Vehicles Act in 2024 but recently have opened the application process for operators to run self-driving passenger services.

Key issues in AV Regulation

These policy and legislative initiatives raise a range of key issues for AV regulation – I want to look at three: safety standards, remote assistants/ operators and Level 3 ADS features:

How safe is safe enough?

How safe should we expect AVs to be? The UNECE draft regulation states that “As a general concept, the safety level of ADS shall be at least to the level of a competent and careful human driver.” And that “The ADS shall be free from unreasonable risk.” Similar language is used in the proposed requirements in China and the US.

The UK is developing a statement of safety principles with the aim under its AV Act 2024 that:

(a) authorised automated vehicles will achieve a level of safety equivalent to, or higher than, that of careful and competent human drivers, and

(b) road safety in Great Britain will be better as a result of the use of authorised automated vehicles on roads than it would otherwise be.

Setting a safety standard based on human driving allows some comparability to existing safety data. It also aligns with some existing standards, such as those for Automatic Lane Keeping Systems.

However, AVs can sense and respond to road conditions in ways vastly exceeding the capabilities of human drivers. Only requiring them to be as safe as a careful and competent /attentive human driver (whilst a minimum standard), seems to me to both lack ambition and to fail to meet public expectations.

It may ultimately be a theoretical point – the AV industry will need to meet very high public expectations on safety (as companies are already claiming to do) to gain acceptance. AV companies will practically need to far exceed the “competent and careful human driver” standard.

Do we need to regulate remote operators/ assistants?

There are various approaches emerging for so-called “remote operators” or remote assistants for AVs. There does seem to be some confusion amongst stakeholders about the role of these workers (confusion reflected in the varying terms used). Feedback from AV operators suggests that they will not be actively controlling vehicles (which would be impractical). Instead they will provide guidance for unusual conditions or scenarios.

The Singapore Consultation Paper raises explicit regulation for this role. But there is a risk in trying to set detailed requirements when operational models are still evolving. Strong safety duties on the companies responsible for AVs, covering actions of their employees including remote assistants, may ultimately be a simpler, more flexible and more effective approach.

Do we need or want Level 3 ADS features?

The UNECE proposed regulations, UK legislation and proposed Chinese and US requirements all include regulation of ADSs operating at Level 3 (conditional automation). These require a fall-back user to be ready to take back control. However, the Singapore government (to its credit) asks in their consultation: “Given the regulatory and legal complications, should we have SAE Level 3 vehicles in Singapore?”.

In addition to legal complications, I would add the potential safety risks of ADSs operating at Level 3. These systems are also likely to only be a transitional technology, with systems moving to Level 4 as they mature. Level 3 features will have many of the costs of Level 4 systems but will have more limited benefits. Governments are examining how they manage the handover of control inherent in these systems, but there do not appear to be agreed answers on how to do this safely. China has proposed to require Level 3 vehicles to be able to bring themselves to a minimal risk condition, moving the requirements closer to those of Level 4 systems.

Given these factors, should governments spend time and money developing regulations for Level 3 ADS features? Or should they instead focus their efforts on Level 4 ADS features?

Where does this leave Australia?

Australia has a longstanding policy of aligning with international vehicle standards so is likely to look to implement the new UNECE regulation in the future. However, this will likely take several years with the UNECE documents yet to be finalised.

To meet the current 2027 timeline for AV regulation in Australia, drafting on Australian legislation could continue, whilst seeking to align the approach with the emerging international regulation. There are also a range of matters outside the draft UNECE regulations that State and Federal AV legislation will still need to cover. These include changes to road rules, compulsory third-party insurance, enforcement and penalties.

It is unlikely that we will see significant deployments in Australia without further legislative support for commercial deployment.

Conclusion

The past six months shows movement towards greater international consensus on key aspects of AV regulation, including safety standards and approval processes.

But there is still significant work for national and sub-national transport agencies to finalise safety requirements; adapt road rules, insurance and operations; and establish appropriately resourced regulators. There is much work to do to develop a comprehensive regulatory framework. The job is far from complete.

Notes:

* “ADS feature of type 1 (ADSF-1)” means an ADS feature which includes an ADS fallback response requiring a fallback user. “ADS feature of type 2 (ADSF-2)” means an ADS feature which does not include an ADS fallback response requiring a fallback user.

**Please note that this analysis is based on a translation of the primary document by Claude, along with secondary sources.