Tesla Rashomon: How far is it from true autonomous driving?

2. Who will take responsibility for autonomous driving? 

"Five innocent people are tied to a tram track, and an out-of-control tram is coming towards them. You can pull a lever to make the tram go to another track. The problem is that there is also a person tied to the other tram track. Will you pull the lever?" 

This is one of the most famous thought experiments in ethics - the "Trolley Problem". This extreme situation rarely occurs in daily driving, but the autonomous driving system must choose between multiple harmful actions. In the future, connected smart cars will be able to obtain data information about nearby people. Should they make selective collisions based on this information before the vehicle loses control? For example, who should make more sacrifices, the elderly or the young? 

The "tunnel dilemma" is an extension of the "trolley problem" - "A self-driving car is carrying passengers and is about to enter a tunnel, when a child suddenly runs out. The car starts to brake, and the software realizes that braking alone will not slow the car down enough for the child to survive, but turning at the same time as braking can sacrifice the safety of the passengers in the car. The car needs to make a decision: continue to drive on the road and hit the child, or turn and hit the wall of the tunnel entrance, causing casualties to the passengers?" 

The question facing self-driving car manufacturers is straightforward: save pedestrians first or save the people in the car first? When Mercedes-Benz said they would save the people in the car instead of pedestrians, the public was outraged and asked Mercedes-Benz to retract the statement. Car companies are thus caught in a dilemma: leaning towards driver safety will cause public outrage; leaning towards public safety will endanger car buyers and cause them not to buy. 

Engineers at MIT crowdfunded a "moral machine" for this purpose, and asked users to make choices in the form of animations on the website to determine people's moral preferences. The experiment collected more than 40 million decision data from 233 countries. These data show that people are more willing to save human lives than other animals; give priority to the statements of young people rather than the elderly; men are more willing to save women's lives... 

In the American TV series "The Upload", passengers can choose whether to "save pedestrians or save themselves".

This series of ethical dilemmas also points to a legal issue - who is responsible for autonomous driving? For L5 cars with only passengers but no real driver, who will bear the driving responsibility? The legal position represented by Germany and Japan believes that the responsible party is the real driver, while the legal position represented by the United States and the United Kingdom believes that the responsible party is the algorithm. In 2017, the European Parliament passed the "Legislative Recommendation to the European Commission on Civil Law Rules for Robots", which believes that robots have a certain degree of autonomous decision-making ability through deep learning, and considers giving intelligent robots the legal status and responsible subject status of "electronic personality" (electronic person). But the problem has not been solved. If it is the algorithm's fault, how to hold the algorithm accountable? 

A series of ethical and legal issues are still under study and discussion, but the commercialization of autonomous driving can no longer wait. In order to keep up with the progress of autonomous driving, more than 20 states in the United States have successively introduced more than 50 bills related to autonomous driving. In 2017, the U.S. House of Representatives unanimously passed the "Automatic Driving Act". The U.S. NHTSA (National Highway Traffic Safety Administration) has successively passed a series of policy guidelines. 

In 2017, Beijing formulated the "Guiding Opinions on Accelerating the Promotion of Road Testing of Autonomous Driving Vehicles in Beijing (Trial)" and the "Implementation Rules for the Management of Road Testing of Autonomous Driving Vehicles in Beijing (Trial)". In 2018, Shanghai issued the "Shanghai Intelligent Connected Vehicle Road Testing Management Measures (Trial)". Subsequently, more than ten cities or regions including Shanghai, Tianjin, Chongqing, Shenzhen, Guangzhou, Changchun, Changsha, Pingtan, Jinan, Hangzhou, and Guangzhou Nansha issued local autonomous driving vehicle testing management regulations. These regulations are mainly aimed at autonomous driving road testing and have a relatively low legal level. 

As autonomous driving technology matures and becomes more commercialized, the ethical and legal issues facing human society will become increasingly complex, including the unemployment of a large number of real drivers, privacy issues caused by highly interconnected information, hacker and malware issues, etc. The high concentration of markets and data will strengthen the monopoly of giant companies, and it is hard to say whether they will not control power beyond the government.   

3. Why should we strive for the leading position in autonomous driving? 

Someone has done the economic calculations for autonomous driving. In the 20th century, about 60 million people died in car accidents, more than the number of deaths during World War II. Even though cars have become safer, more than 1.4 million people still die in traffic accidents every year around the world. Telematics company Inrix estimates that traffic congestion in the United States causes economic losses of more than $305 billion each year. Nearly one-sixth of the workforce is restricted from traveling due to driving ability and qualification issues. People are pinning their hopes on autonomous vehicles based on the Internet of Vehicles to solve these problems. 

In the era of autonomous driving, shared travel is expected to become the main mode of travel. Research institutions predict that new car sales will drop by about 50%. Of the approximately 30 million autonomous vehicles produced each year, half will be sold to China, another quarter to the United States, and the rest will be scattered in the European Union, Japan and emerging markets. The global automobile manufacturing industry, worth $2 trillion, will be incorporated into the larger personal transportation service market. This market is expected to reach a scale of $7 trillion to $10 trillion by the middle of this century, roughly equivalent to the economic size of the entire European Union today. 

Governments around the world are actively clearing policy and regulatory barriers for autonomous driving. The United States emphasizes that autonomous vehicles, as part of the transportation system, should be deeply integrated into the existing transportation system. The European Union emphasizes vehicle-road collaboration and European integration. Japan attaches importance to the implementation and industrialization of autonomous vehicles. 

"Autonomous Driving 4.0" released by the United States in 2020 proposed integrating the work of 38 federal departments, independent agencies, committees and the Presidential Executive Office in the field of autonomous driving to ensure the United States' leading position in the field of autonomous driving technology. 

Since 2016, the U.S. Highway Safety Administration has been advancing relevant bills every year.

The European Union released the "Road to Automation: Europe's Future Mobility Strategy", planning to achieve a car networking model in which all new cars are equipped with communication functions by 2022, and to enter a society with fully autonomous driving as the standard by 2030, with the goal of making Europe a world leader in the field of fully autonomous driving. Japan's "Report and Plan 4.0 on the Implementation of Autonomous Driving" published in May 2020 proposed to expand unmanned autonomous driving services to 40 regions across the country by 2025. 

In 2017, the State Council issued the "New Generation Artificial Intelligence Development Plan", and autonomous driving was included in the national strategy. In February 2020, the National Development and Reform Commission and 11 other departments jointly issued the "Intelligent Vehicle Innovation and Development Strategy", proposing a strategic vision for 2025, and the technical innovation, industrial ecology, infrastructure, regulations and standards, product supervision and network security system of standard intelligent vehicles have basically taken shape. The "National Comprehensive Three-dimensional Transportation Network Planning Outline" issued this year proposed that by 2035, the total scale of the physical line network of the national comprehensive three-dimensional transportation network will be about 700,000 kilometers. 

Since 2015, China has launched a series of planning plans for autonomous driving.

In the "Autonomous Driving Vehicle Maturity Index" released by KPMG in 2020, China ranked only 20th. The main factors affecting China's ranking are policy legislation and infrastructure.

The Autonomous Vehicle Maturity Index assesses the deployment and readiness of autonomous driving in a total of 30 countries and regions around the world, and the rating ranking is broken down into four dimensions: policy and legislation, technology and innovation, infrastructure, and consumer acceptance. 

KPMG Autonomous Vehicle Maturity Index

There are two paths for the development of global autonomous driving technology: one is the driving automation of traditional cars, that is, the innovation and application of autonomous driving technology by traditional automobile manufacturing powers; the other is to graft Internet technology into cars, that is, to innovate and apply artificial intelligence learning, perception, and control technology to driving technology. The automobile faction, such as GM, Audi, BMW, Toyota and other traditional automobile manufacturers, follows the order of L0, L1, and L2. Most of the Internet faction directly cuts in from L4 and grafts autonomous driving technology onto cars.