In our previous article, we touched upon the promising potential of connected cars as well as the related developing legal frameworks. For the third article of the series, we will focus on the "A" of CASE - Autonomous vehicle, one of the hottest topics in the industry that promises a game-changing feature for tomorrow's cars that will forever the change the future of mobility.

Autonomous cars in a nutshell

The idea of making self-driving vehicles emerged almost a century ago but has never been closer to reality as we have seen in recent years. To automate a vehicle, engineers install electronic systems called "advanced driver-assistance systems" (ADAS) to equip vehicles with "eyes and brains". ADAS, which consists of numerous sensors "the eyes" and a sophisticated AI "the brains", as commercially available today, has brought us half-way towards the ultimate goal of fully-autonomous cars. These features include an adaptive cruise control system that can maintain a chosen velocity and distance between a vehicle and the vehicle ahead, thus allowing additional time for drivers to take a rest from the acceleration and brake pedals. Other key highlights of ADAS also include emergency brake assistance, lane keeping and automatic parking that have proven to help drivers drive more safely and comfortably. To achieve fully automated vehicles that can drive under all conditions, automakers, auto-part manufacturers, system developers, academic organizations and many others are now rushing to develop ADAS and AI for various dynamic driving tasks.

The most widely recognized and adopted categorization standard of autonomous vehicles has been introduced by SAE International (previously known as the Society of Automotive Engineers), a US-based professional association and standard developing organization. Under SAE taxonomy and definitions, driving automation ranges from no driving automation (Level 0) to full driving automation (Level 5). Technically speaking, automated driving features in the world of autonomous vehicles start from Level 3. In contrast, vehicles of Level 2 or below only provide driver assistance features. Level 5 autonomous cars are defined as vehicles which can be driven as completely driverless cars everywhere in all conditions, i.e. the "true full-autonomous car".

Since it can be foreseen that autonomous cars will lead to an increasing use of radio equipment for vehicle centered communications (commonly known as vehicle-to-everything or V2X), arrangement of radio spectrum band and standards to serve these sensing, transmissions and communications requires both technical and legal developments. In Thailand, as mentioned in our first article of this series, several regulations have been amended and will be promulgated to support the change.

AI law in Thailand is still in the initial stages of development but has been making consistent progress. As the development of AI has been included in several national strategic plans, earlier this year, the Ministry of Higher Education, Science, Research and Innovation, in collaboration with the Ministry of Digital Economy and Society, set up a task force to work on the drafting of Thailand's national AI model plan for further approval from the Cabinet. More government agencies are expected to get involved in future AI initiatives.

Needless to say, when vehicles have to deal with more data processing and communications, cybersecurity and data privacy related regulations are always crucial topics to be concerned of.

Driving Safety - an uncompromising priority

In the automotive industry, nothing is more critical than safety. Technologies for autonomous cars therefore require the highest attention under this topic. However, no matter how many tests have been made to confirm the safety of a new system, even the slightest hint of a potential safety issue can damage consumers' trust.

In Thailand, road accidents have long been a chronic issue. According to a World Health Organization report[1], the rate of road fatalities in Thailand remained practically constant over the last ten years, placing the country as worst in ASEAN and among the worst globally. A recent joint research conducted by the Thailand Accident Research Center and automotive companies in Thailand reveals that more than 90% of road accidents are attributed to human driving behavior and not vehicle failure, road conditions or the environment.

Driving safely requires good control of a vehicle and good real-time judgment. Some studies divide driver-related factors which may result in crashes into five groups: (1) sensing and perceiving (e.g. driver distraction and failure in recognizing incoming hazard), (2) predicting (e.g. misestimating the velocity of the other vehicles), (3) decision making (e.g. to brake, accelerate or turn), (4) execution and performance in controlling the vehicle, and (5) incapacitation (e.g. drug or alcohol use, poor health conditions or drowsiness). Leaving aside the event of failure in an auto-driving system, with autonomous cars equipped with accurate sensing devices and refined AI, most crashes, if not all, caused by the above human errors can be prevented. At the least, we can rest assured that the AI driver would always be "awake and sober" during the trip.

An aging society is also another key factor to bring autonomous cars on the road. Driving plays an important role in maintaining an elder's independence and as there is no upper age limit for driving cars, an elderly individual that is suffering from deteriorating vision and slower reflexes may cause more risks on the road. This is where autonomous featured vehicles may fill the gap. Automated cars will also provide those who have specific needs, such as pregnant women or the disabled, with more convenient options to travel.

Considering the fact that a complete risk-free system does not exist, balancing risks by taking into consideration the facts, the rules and the currently available tools should lead to the optimum solution. In this sense, leveraging autonomous car technologies will definitely become a part of driving safety solutions. Nonetheless, realizing autonomous vehicles on the road will be highly dependent on local government's policies and, above all, trusts from consumers.

Regulations do matter

While the development of automated cars is moving at a fast pace, it seems that the biggest hurdle in bringing them to reality is the development of relevant legal frameworks. Authorities in many jurisdictions are aware of the facts that the technologies are out there waiting only for official recognition to reflect them into local laws. There have been several promising trends which give a strong push to add autonomous cars on real roads in several jurisdictions. Below are some noteworthy milestones of autonomous car related legislation.

  • US: In July 2019, the California Public Utilities Commission granted Waymo, a US autonomous driving technology development company, permission to commercially operate driverless vehicles as a carrier of passengers, a giant leap towards opening up the regulatory pathway for the operation of driverless vehicles. Waymo's autonomous vehicles are Level 4, i.e. no human driver is needed in specifically defined operational conditions. To date, Waymo is the world's only one ride hailing service provider using autonomous cars.
  • EU: In 2016, EU member states, the European Commission and the private sector agreed in the Declaration of Amsterdam to join goals and joint development in automated and connected driving technologies. While a legal framework for autonomous vehicle approval in the EU is still developing, in April 2019, the Technical Committee - Motor vehicles (TCMV) of the Commission published Guidelines on the exemption procedure for the EU approval of autonomous vehicles. The guidelines, which focus on Levels 3 and 4 vehicles, are aimed to harmonize the practice of autonomous vehicles assessment of each member state as well as to ensure fair competition and transparency.
  • Japan: Following the issuance of the official guideline for autonomous cars in 2018, in April 2020, several vehicle and traffic related laws including, among others, Road Trucking Vehicle Act and Road Traffic Act,  were amended to allow the use of vehicles equipped with "automated driving apparatus", which must meet the requirements as prescribed by the Minister of Land, Infrastructure and Transportation and must be equipped with a drive-recording device. The competent organization which has the authority to regulate and approve the automatic driving apparatuses and their AI programs has also been established. These developments have paved the way for Japan to gain momentum of the actual use of (specifically, Level 3) autonomous cars commercially available in the country.

Among these positive advancements to support the practical use of autonomous cars, questions remain on the fundamental, though principal, issue -- liability. Given the fact that, while in the automated driving mode, autonomous cars shall drive in accordance with the traffic rules, have predictable and careful behavior and hence shall not cause traffic accidents that are rationally foreseeable and preventable, to a large extent, autonomous cars should theoretically eliminate conventional road accidents caused by violation of traffic rules. However, as long as human-driven cars (Levels 0 to 2) and autonomous cars (Levels 3 to 5 in automated driving mode) co-exist, the risks remain.

In Thailand, tort principles apply when damages are caused intentionally or by negligence of the drivers. However, when the operational control of driving become more vague with automated features (especially in Level 3 autonomous cars), legal frameworks will play a vital role in clarifying whether the civil liabilities fall on the drivers, the vehicle makers, the system developers or any other related parties. Other existing liability regimes such as product liability law were also not originally developed to deal with these challenges. To address these issues, it may require entirely new concepts tailor-made for autonomous vehicles. The direction of this legal development is expected to disrupt many industries, specifically the auto insurance market.

As to criminal liability, where intent is a core element to establish culpability, the current legal framework of each jurisdiction may require further review and amendments to identify and clarify criminal liability, taking into account the nature and degree of autonomous driving. Legally, under the current universal principle of criminal law, if the vehicle is fully automated and the person in the vehicle does not conduct any acts to take over the driving control, should an accident occur, such person would not be subject to any criminal liabilities since the pure intent of action is only to transport from one place to another using the autonomous vehicle. However, proving that the accident occurred while the vehicle is automated or not can be technical and difficult to establish. As fully-automated vehicles become available the regulators in each jurisdiction will need to catch up with these challenges. Even in the jurisdictions which already allow the use of autonomous vehicles on the roads, many of their legal frameworks are still silent on this area.

The liability issues could be even more complex when factual circumstances are involved. The automated driving system which has not been fully installed with the most updated data (e.g. virtual maps and local traffic regulations, which may change from time to time) can cause issues. In more critical supposition, the actual environment may even differ from what is reflected in the latest data set.

The development of the legal framework in relation to the concept of liabilities arising from the use of autonomous vehicles will be one of the decisive elements for successful deployment of this game-changing technology where the middle ground of all stakeholders may differ in each jurisdiction.

New exciting possibilities await

The largest benefit of autonomous vehicles may be safety, but there are more business opportunities and disruptions to come. In the not too distant future, with fully automated driving features, vehicle owners may just program and "rent out" their vehicles to pick up and drop off packages or passengers at designated places while having the "AI drivers" do their jobs. Many industries, such as construction and agriculture businesses, will also benefit from smart and precise automated features. More creative and disruptive businesses using autonomous vehicles will emerge for better comfort and reassurance of hygienic and disease preventive measures in the world post-COVID. Again, legal frameworks hold the key for the next-generation vehicles, and all of us, as stakeholders, need to be prepared for the coming changes.

In the next article, we will discuss the "S" of CASE - Shared Mobility - in more detail.