In February 2024, Elon Musk surprised investors with a post on X announcing that Tesla would unveil a robotaxi in August the same year.
However, just a few weeks ago, he announced that this event had been postponed until October due to design change.
Tesla isn’t the only carmaker facing difficulties in its quest for complete autonomy. Other companies, including Cruise, a subsidiary of General Motors, and Waymo, a subsidiary of Alphabet, the parent company of Google, have encountered setbacks in the race toward autonomous cars.
In October last year, Cruise had its operating license suspended in California after one of its driverless cars hit a pedestrian and dragged him for 6 meters before coming to a halt. In February, Waymo announced it was voluntarily recalling the software for its autonomous robotaxis after two of its vehicles collided with the same recovery truck within minutes of each other.
Despite these challenges, companies remain committed to solving the problem of autonomous driving. In Europe, incumbent carmakers such as BMW and Volkswagen have active autonomous vehicle development programmes. Wayve, a British start-up, is a leader in next-generation autonomous car technology. In China, leading companies in this field include BYD, a major player in the electric vehicle sector, and the Apollo business unit of technology giant Baidu.
An ambitious vision with great challenges
To understand the challenges of creating fully autonomous vehicles, it’s essential to know how they operate.
Autonomous driving generally follows a “detect-plan-act” architecture. Vehicles use a suite of sensors (such as LiDAR, radar, and cameras) and sophisticated AI software to process sensory data in real-time. This system decides on an action in a fraction of a second and sends signals to the car’s steering, braking, and acceleration control systems.
Each manufacturer has its own approach. For example, Tesla has avoided using the more expensive LiDAR technology, instead relying on cameras, ultrasonic sensors, and radar. Meanwhile, Wayve’s AV2.0 autonomous driving technology focuses on learning and adapting to new scenarios without requiring pre-programmed rules or detailed maps.
The industry has made significant progress in bringing autonomy to market. Many commercially available cars now include autonomous features such as driver assistance and advanced cruise control. Companies like Waymo and Cruise are testing Level 4 autonomy, which offers full autonomy but is currently limited to specific service areas and scenarios.
Technological, regulatory and ethical challenges
The ultimate goal is to produce Level 5 autonomous vehicles, capable of driving themselves on any road, in any conditions, without any passenger intervention. Developing such vehicles involves numerous technological, regulatory, and ethical challenges.
Technologically, these vehicles use sensors like LiDAR to perceive the environment in real-time. However, the reliability of these sensors can decrease with an increase in the number of vehicles on the road. Additionally, the machine learning algorithms that train the AI systems require vast amounts of data covering all possible driving scenarios. While today’s autonomous cars can handle many standard situations, unexpected ‘extreme’ cases pose significant challenges for their systems.
Ethically, there are complex questions such as the “trolley problem” or “moral dilemma.” How should autonomous cars make critical decisions that involve people’s safety? This dilemma goes beyond programming challenges and involves profound moral decisions traditionally made by humans. Who should define these ethical frameworks—engineers, regulators, or someone else?
Moreover, there are significant concerns about cybersecurity, as autonomous cars could be vulnerable to hacking. Given these challenges, there are no simple solutions; ongoing dialogue between governments, the private sector, and the public is crucial to navigating these complexities.
Towards major change
The advent of autonomous cars is set to significantly disrupt the traditional automotive economy. This includes the car insurance industry, given the likelihood a shift of legal responsibility from drivers to manufacturers and technology developers. Additionally, the traditional business model of vehicle ownership may evolve towards car-sharing and subscription-based models, potentially reducing the total number of cars on the road.
In terms of competitiveness, companies that successfully deploy AI and software for safe autonomous driving could gain a significant advantage over traditional manufacturers. Thus, autonomous cars represent not just a technological evolution but also an economic revolution in the automotive sector.
When will we see a Level 5 autonomous car?
Although the latest forecasts from S&P Global Mobility predict that Level 5 passenger cars will not appear before 2035, there is already targeted use of Level 4 autonomous vehicles. A McKinsey study predicts that Level 4 autonomous driving systems could generate sales of between $170bn (£129bn) and $230bn (£174bn) by 2035.
Many experts believe that long-haul trucks will be the first to adopt autonomous technology on a large scale. Daimler Truck recently unveiled its first autonomous truck, announcing that driverless semi-trailers could be on the road by 2027.
Consumer acceptance is crucial for the widespread adoption of autonomous vehicles. A survey by S&P Global Mobility shows that 65% of US buyers are interested in Level 2+ automated driving on highways, but only 25% show strong interest in advanced autonomous driving features – although this lukewarm interest may increase over time.
For the foreseeable future, competition will be intense, with the US, China, Germany, and Japan heavily investing in this technology race. Both new and established companies are continuously innovating around key technologies, business models, and use cases for autonomous vehicles.
This article was authored by Michael Lenox, Professor of Business Administration at the Darden School of Business, University of Virginia.
