Autonomous vehicles

Cover: Image generated using Dall-E integrated into Bing.

20 min. read

Maurício Pinheiro

Introduction

The automotive sector is undergoing a revolution as autonomous vehicles (AV) transform how we travel. These automated vehicles can recognize their surroundings, make judgments, and perform maneuvers without the assistance of a human driver since they are controlled by computers and artificial intelligence (AI). The degree of automation in self-driving cars is always rising. The benefits of this technology include increased user convenience, cost savings, and greater driving safety. However, there are drawbacks and moral dilemmas associated with their broad use, such as worries about security and the requirement to make moral judgments in emergency situations. The many levels of automation, the benefits and difficulties of autonomous vehicles, potential moral ramifications, and societal changes will all be covered in this article.

What is an autonomous vehicle?

Autonomous vehicles (AV), also commonly known as self-driving cars, represent a groundbreaking technological advancement with the ability to operate and navigate without human intervention. These vehicles utilize sophisticated computer control systems that drive them, eliminating the need for human drivers. While it typically takes humans around forty-five hours of training to become proficient in driving, which entails mastering complex skills such as perception, navigation, decision-making, and vehicle control, autonomous vehicles rely on a combination of Artificial Intelligence (AI), neural networks, and mechanical components to perform these tasks.

The core of AV technology resides in the integration of AI systems, powered by advanced neural networks. These neural networks are designed to process vast amounts of data gathered by the vehicle’s sensors and make real-time sense of it. Cameras, LiDAR (Light Detection and Ranging), and radar sensors are deployed to endow the vehicle with a comprehensive perception of its surroundings. Working in unison, these sensors detect and identify objects, pedestrians, road markings, and other vehicles, allowing the vehicle to comprehend its operating environment.

In addition to perception, autonomous vehicles rely on GPS (Global Positioning System) for precise navigation. By utilizing satellite signals, AVs can determine their exact location, chart the desired route, and make necessary adjustments during the journey. The amalgamation of sensor data and GPS information enables these vehicles to develop a detailed understanding of their present position and effectively plan their movements on the road.

The AI algorithms employed in AVs play a vital role in inference, which involves interpreting and analyzing sensor data to make informed decisions. These algorithms process the copious amounts of information received from the sensors and analyze it to determine the appropriate course of action. For instance, the algorithms can identify road obstacles or hazards and respond accordingly by adjusting speed, changing lanes, or applying the brakes. The continuous integration of sensor data and AI algorithms empowers autonomous vehicles to make split-second decisions in intricate and dynamic driving scenarios.

The decision-making and planning capabilities of autonomous vehicles rely on a combination of specialized rules and statistical estimates. The AI algorithms take into account various factors such as traffic conditions, road regulations, weather conditions, and the vehicle’s own capabilities to make decisions that prioritize safety, efficiency, and passenger comfort. Through the continuous analysis and updating of this information, AVs can adapt to changing conditions and make optimal real-time decisions.

The technology of self-driving cars is always changing, going through several stages of automation. The Society of Automotive Engineers (SAE International) categorizes these levels as follows, from “Level 0” to “Level 5”:

• Level 0 (No automation): All driving activities are performed by the human driver, while AI monitors the road and warns the driver when necessary.
• Level 1 (Driver Assistance): AI can only undertake a certain duty, such as steering control, when prompted by the human driver.
• Level 2 (Partial Automation): AI can undertake numerous functions (such as steering, braking, and accelerating), but the human driver must still observe and take control when necessary.
• Level 3 (Conditional Automation): AI can take over driving, but the human driver must be ready to regain control when the AI requests it. Some question whether this abrupt change in control will increase danger rather than reduce it.
• Level 4 (High Automation): AI can completely take over driving.

The current stages of autonomous vehicle technology differ across the globe. Several firms, including Tesla, Waymo (from Google), Uber, and several traditional manufacturers such as General Motors, Ford, and BMW, have invested in research and development in this field. AVs in various stages of development and implementation include the Tesla Model S, Waymo One, Cruise Origin, and BMW iNEXT.

Society will be transformed as autonomous car technology improves. The future of transportation will be heavily influenced by Level 4 autonomy, in which buses and trucks operate on defined routes, and Level 5 autonomy, in which robot taxis are summoned via applications like Uber. Autonomous driving will become more common as technology advances and interest grows, providing increased safety, efficiency, and convenience.

Implications of fully AVs (Level 5):

Autonomous vehicles will usher in a transportation revolution by providing convenient, cost-effective, and safer on-demand solutions. Uber will, for example, efficiently orchestrate its fleet of autonomous vehicles using complex AI algorithms, strategically deploying them near individuals who may soon require transportation, such as after a show, or a concert. This approach improves overall efficiency by optimizing vehicle routes to minimize wait times and including battery recharges along the way.

This fully automated and AI-managed fleet optimizes vehicle usage while decreasing uncertainty in the absence of human drivers. As self-driving cars become the standard in ride-sharing services, the huge cost reduction, which presently mostly benefits the driver, will make journeys significantly more reasonable. This affordability issue will encourage people to rely less on personal automobiles.

In the long run, autonomous vehicles are expected to outperform human drivers in terms of safety. While good human drivers normally accumulate roughly ten thousand hours of driving experience, autonomous vehicles have the potential to accumulate an incredible amount of hours, potentially in the billions. This is accomplished by their ability to learn continuously, in which each autonomous vehicle learns from the experiences and encounters of all other autonomous vehicles. AVs can enhance their driving performance and safety measures considerably over time by constantly analyzing and assimilating this massive amount of data.

However, governments and regulatory agencies will only grant broad permission for the widespread deployment of AVs if they can demonstrate that they are “safer than humans.” With an estimated 1.35 million highway fatalities worldwide each year, any deployment of AI technologies must meet, if not exceed, the safety standards achieved by human drivers. This is critical to ensuring public trust and acceptance of autonomous vehicles as a safe and dependable means of transportation.

Extensive testing and thorough assessment will be required to demonstrate the improved safety features and performance of self-driving automobiles. These tests will include rigorous simulations, controlled experiments, and real-world trials to evaluate the efficiency of different autonomous driving systems, sensors, and AI algorithms in preventing accidents and managing possible dangers.

Furthermore, legislative frameworks and industry standards will be developed to govern autonomous vehicle safety needs. These frameworks will define benchmarks that autonomous car manufacturers and service providers must meet and exceed. Audits, inspections, and certification processes will be undertaken on a regular basis to verify compliance with safety rules, encourage accountability, and preserve public faith in the technology.

The progress of autonomous vehicle technology is set to cause a major shift in many aspects of our society. While the major advantages will likely be recognized in terms of lives saved and enhanced productivity, broad adoption of autonomous cars will also disrupt countless businesses and redefine the employment economy as we know it.

Taxi drivers, truck drivers, bus drivers, and delivery drivers will face the most substantial interruptions. These occupations will experience severe unemployment in a world where autonomous cars rule the roads. Currently, almost 4 million Americans rely on driving trucks or taxis as their major source of income, not to mention the innumerable part-time drivers who work for businesses (such as Uber), postal services, delivery services, warehouses, and more. These occupations, which have been essential to our transportation and logistics sectors, will be gradually displaced by AI systems that can do these operations autonomously and with efficiency and precision.

In addition to that, the disruptions generated by autonomous cars will have an influence on the rearrangement of other traditional professions in addition to causing employment losses. For example, car maintenance will experience a fundamental transformation. While mechanical repairs have traditionally been the major emphasis of the automobile business, the emergence of self-driving cars will entail a shift toward competence in electronics and software. To handle the sophisticated technical components and powerful AI systems embedded in these cars, technicians and mechanics will need to change their skill sets.

The implications of self-driving cars will also have an impact on sectors involved with existing transportation infrastructure. Gas stations, for example, will witness a significant fall in demand since the bulk of autonomous cars will be powered by electricity. Car dealerships will also undergo restructuring as car purchase and ownership patterns shift. Furthermore, as autonomous cars may be shared, parked remotely, or operated constantly without requiring lengthy periods of stationary parking, the demand for parking lots and garages will be minimized. As a result, in order to stay relevant and sustainable in the age of autonomous mobility, these businesses will need to adapt to the changing terrain and explore new routes.

This transformation’s societal impact will extend beyond job losses and economic disruptions. Many lives will be irrevocably affected, similar to the societal upheavals witnessed with the move from horse-drawn carriages to cars. People will have to acclimate to a new environment in which human-driven cars are less common and autonomous vehicles are the norm. To enable the safe and effective integration of autonomous cars into our everyday lives, legislation, insurance policies, and infrastructure development will need to be adjusted. Driving an automobile in the future will be more like a pastime than an exercise.

Ethics in Autonomous Cars

Several problems, including ethical concerns, must be overcome before autonomous vehicles become mainstream. This is to be anticipated because millions of lives, not to mention numerous industries and hundreds of millions of jobs, are at risk. There will be times when autonomous cars must make difficult ethical judgments.

The traditional “trolley problem,” which boils down to a decision between acting and killing person A or not acting and allowing individuals B and C to die, is perhaps the most renowned ethical quandary in this subject. What if person A is a youngster, and you think the solution is obvious? What if individual A is your child? What if the automobile is yours and person A is your passenger?

When human drivers cause deaths, a legal procedure is in place to examine whether they performed legally and, if not, to impose the necessary sanctions. However, as we move into an era in which AI systems make decisions, the question of what happens if AI causes a fatality emerges. Can AI produce human-understandable reasons for its judgments that can be legally and morally justified? This challenge introduces the notion of “Explainable AI” (XAI), which refers to AI systems’ capacity to deliver visible and interpretable explanations for their activities.

Because AI systems are taught on data and their replies are the product of elaborate mathematical equations, achieving XAI is a difficult undertaking. These complicated algorithms must be greatly reduced in order for people to grasp their decisions. This simplification approach is difficult since it necessitates balancing accuracy, interpretability, and the basic substance of the AI’s decision-making process.

The discussion of ethics in Artificial Intelligence (AI), however, has generated misunderstandings, false problems, and misleading perspectives. Often, debates on the subject are marked by biases and misguided questions, leading to false solutions. An example of this appeared in the article titled “The Moral Machine,” published in the renowned journal Nature in October 2018. In that article, a research study was presented, claiming to be “the largest ever conducted on machine ethics.” Over 18 months, a group of researchers interviewed 2.3 million people in various countries, collecting about 40 million responses regarding hypothetical situations involving accidents caused by autonomous vehicles.

The research divided the participants into three groups: one composed of North American and European countries with Christian tradition, another including Asian countries with Confucian and Islamic traditions, and a third composed of South and Central American countries, as well as former French colonies in Africa. However, the justification for this division appears arbitrary, raising questions about criteria such as religion, ethnicity, and nationality.

The questions posed to the participants were predictable, and the responses followed expected patterns. For example, the majority agreed that an autonomous vehicle should prioritize human life over animals and that, in a situation where a choice had to be made between saving a group of people or an individual, the individual should be sacrificed. However, some questions seemed unnecessary or biased, such as the one involving running over an executive or a homeless person. Furthermore, certain important issues were not addressed, such as the choice between saving a woman or a man, or a person of color compared to a white person.

An important conclusion from this research is that an algorithm of an autonomous vehicle should take into consideration a wide range of data and moral preferences before making a decision in risky situations. This raises questions about the technical feasibility of obtaining and processing such data in real-time. Additionally, the research reveals that the problem lies not only in the algorithm itself but in human nature, as it is humans who program these algorithms and conduct research on ethics in AI. And their moral preferences vary according to the demographic and cultural characteristics of those involved.

Final reflections

In conclusion, autonomous vehicles represent a revolution in transportation, offering greater convenience, enhanced safety, and cost reduction. With advancing technology and increasing interest from companies and consumers, autonomous driving is becoming increasingly common. As level 5 autonomous cars hit the roads, it is expected that ride-sharing services will be widely adopted, reducing transportation costs and encouraging people to relinquish ownership of their own vehicles. Furthermore, in the long term, autonomous vehicles have the potential to significantly improve traffic safety as they can accumulate trillions of hours of experience. Although there are ethical challenges to be addressed, such as decision-making in risky situations, it is possible for the technology to evolve to ensure the safety and justifiability of autonomous vehicle actions. With reduced traffic congestion, more efficient vehicle usage, and decreased air pollution, society as a whole is expected to benefit from these technological advancements.

---

Here are some data on autonomous vehicles:

1. 55% of small businesses believe they will have a fully autonomous fleet within the next two decades. Source: Nissan (2018)
2. There are over 1,400 autonomous vehicles in the US, tested by more than 80 companies. Source: TechCrunch (2019)
3. Only 57% of people familiar with autonomous cars would be willing to travel in them. Source: US News (2018)
4. 75% of people prefer driving a car over riding in an autonomous vehicle. Source: Advocates for Highway and Auto Safety (2019)
5. The global autonomous vehicle market is currently valued at $54 billion. Source: Vox (2019)
6. In 1977, the Tsukuba Mechanical Engineering Laboratory in Japan introduced the first semi-autonomous car. Source: Wired (2017)
7. Google’s autonomous car was involved in its first accident in February 2016. Source: Wired (2016)
8. A Honda sedan collided with a Waymo autonomous van, resulting in minor injuries. Source: USA Today (2018)
9. The Tesla Model S caused its first death in Florida while in autopilot mode. Source: The Guardian (2016)
10. The latest Tesla safety report shows that its autopilot is safer than a human driver. Source: Futurism (2020)
11. The first fatal accident involving an Uber autonomous car occurred in Tempe, Arizona, due to human error. Source: The Verge (2017)
12. Uber’s autonomous vehicles were involved in 37 accidents prior to the fatal Arizona crash. Source: Business Insider (2019)
13. Two years after the Uber car’s fatal accident, Uber’s autonomous vehicles were allowed back on California roads. Source: BBC (2020)
14. Elaine Herzberg was the first pedestrian killed by an autonomous car. Source: The Verge (2018)
15. Human-driven cars are considered a fairly safe activity, with one death per 100 million miles driven in the US. Source: Vox (2019)
16. Autonomous cars are designed to include an emergency stop button. Source: The Conversation US, Inc. (2020)
17. 16% of residents would feel comfortable letting a fully autonomous vehicle drive without any control. Source: GovTech (2020)
18. 57% of people would not feel comfortable buying an autonomous car, even if money were not a problem. Source: Driverless Media (2020)
19. Half of women in the US believe that life and death decisions cannot be taught to any type of vehicle, while two-thirds of men agree. Source: Robotics Business Review (2020)
20. The majority of US drivers (71%) feel uncomfortable sharing the road with autonomous vehicles. Source: Gallup (2018)
21. About 94% of traffic accidents are caused by human error, and autonomous technology is expected to significantly reduce these numbers. Source: National Highway Traffic Safety Administration (2015)
22. Widespread implementation of autonomous vehicles may result in a 90% reduction in traffic-related deaths. Source: McKinsey & Company (2015)
23. The main barrier to the adoption of autonomous vehicles is concern for safety and public trust. Source: Brookings Institution (2018)
24. Autonomous technology has the potential to improve mobility for the elderly, people with disabilities, and rural populations. Source: US Department of Transportation (2018)
25. Autonomous vehicles can offer economic benefits such as reduced transportation costs and increased energy efficiency. Source: International Transport Forum (2015)
26. Regulation and legislation surrounding autonomous vehicles vary from country to country, and many governments are working to update existing laws to accommodate this emerging technology. Source: World Economic Forum (2018)
27. The global autonomous car market was valued at over $27 billion in 2021. The market is expected to grow in the coming years, reaching a size of nearly $62 billion by 2026. Source: Statista (2022)
28. China is the largest autonomous car market in the world, with a 45% market share in 2020. China also leads in terms of testing and deployment of autonomous cars, with over 200 companies and 60 cities involved. Source: Statista (2021)
29. Tesla launched its Full Self-Driving (FSD) beta system in October 2020, allowing selected owners to test the feature that enables the car to drive itself in almost all situations. However, the system still requires human supervision and has faced criticism for its safety and accuracy. Source: The Verge (2021)
30. Waymo, a subsidiary of Alphabet, became the first company to offer a commercial autonomous taxi service in the US in October 2020. The service, called Waymo One, operates in Phoenix, Arizona, and allows users to request a driverless autonomous car through an app. Source: BBC (2020)
31. Apple is reportedly working on its own autonomous car project, known as Project Titan. Although the company has not officially confirmed its plans, rumors suggest that it may launch its first autonomous car in 2024 or 2025, using its own battery and sensor technology. Source: Reuters (2020)

---

#AI #AutonomousVehicles #AutomotiveTechnology #ArtificialIntelligence #Automation #TrafficSafety #TransportEfficiency #Mobility #Innovation #AutomotiveFuture #EthicalChallenges #SocialChanges #WidespreadAdoption #AI #AutonomousDriving #Tesla

References:

1. The Holy Driver, chapter 6. in Lee, Kai-Fu, and Chen Qiufan. AI 2041: Ten visions for our future. Currency, 2021.
2. Awad, E., Dsouza, S., Kim, R. et al. The Moral Machine Experiment. Nature 563, 59–64 (2018).
3. Coelho T. Inteligência Artificial, ética artificial, p. 65 in Inteligência Artificial: avanços e tendências. Organizadores: Fabio G Cozman, Guilherme Ari Ponzi, Hugo Neri. São Paulo : Instituto de Estudos Avançados, 2021.

---

---

Copyright 2026 AI-Talks.org