Autonomous vehicle ethics: Planning for safety and accountability

• Author:
• Author name

  Quantumrun Foresight
• April 11, 2023

Autonomous vehicles use software to determine their course to minimize the impact of collisions. These vehicles are designed to reduce the risk of crashes by constantly monitoring their surroundings, predicting potential hazards, and adjusting their course of action accordingly. However, as these vehicles become more advanced, machine judgment is causing ethical dilemmas and public concerns about their safety. 

Autonomous vehicle ethics context

Stakeholders have different expectations of autonomous vehicles: users expect efficiency and reliability, bystanders expect to be safe, and the government expects transportation efficiencies. Backed by years of research, 360-degree vision and sensors, and better information processing power than humans, such vehicles assign risk weightings to situations and make quick decisions for the supposedly best course of action. It has been argued that the intelligence behind the technology will make better and faster decisions than humans in the case of collisions.

The question remains as to who will be at fault when a collision does occur. Is it okay for artificial intelligence (AI) to choose which lives to value and which to save when faced with the option? Germany proposed that such cars should always aim to minimize fatalities and value human life without discrimination. This proposal resulted in mixed views about how much the government should be able to place a value on life. Moreover, it has been argued that the technology is based on the ethics of the engineers who designed it. Some say arbitrary decisions are better than predetermined programs determining casualties. The possibility of autonomous vehicles being hacked or malfunctioning further adds to ethical dilemmas. 

Disruptive impact 

Ethical concerns surrounding fully automated cars include issues such as how the vehicle will make decisions in emergencies, who will be held responsible in an accident, and how to ensure that the car's programming does not discriminate against certain groups of people. These concerns may cause some individuals to be hesitant about switching to fully automated vehicles and may also lead to increased pressure for product engineers to be more transparent about the algorithms used in the cars.

One potential solution to these ethical concerns is mandatory requirements for automated black boxes, which can help to determine the cause of accidents. However, government intervention in this area may also be met with resistance, as some may argue that it is not the government's role to regulate the use of autonomous vehicles. 

Insurance companies will also have to adapt to the advent of fully automated cars. They will need to redesign their policies to account for these vehicles' unique risks and liabilities. These plans may include preparing for instances of product malfunction and determining who will be held responsible in the event of an accident. Comprehensive protection is necessary since there have already been incidents of autonomous car systems misidentifying pedestrians as objects, leading to accidents.

Implications of autonomous vehicle ethics

Wider implications of autonomous vehicle ethics may include:

• Increasing public mistrust of autonomous vehicles, especially if manufacturers are not transparent about their AI ethical guidelines.
• Regulatory bodies requiring autonomous car manufacturers to publish their AI policies and resiliency plans for errors caused by these systems.
• Insurance firms creating comprehensive plans that deal with AI-related faulty systems and cyber hacking.
• With the rise of autonomous vehicles, people's data may be collected and shared with third parties without their knowledge or consent.
• The shift to autonomous vehicles could lead to job losses for human drivers but also create new jobs in areas such as vehicle maintenance, data analysis, and dispute management.
• Potential discrimination against certain groups of pedestrians, particularly if the training data is biased.
• Autonomous vehicles are vulnerable to hacking and cyber-attacks, which could compromise the safety of passengers and other road users.

Questions to consider

• Would you trust an autonomous car as a passenger or a bystander?
• Do you believe public fears would slowly dissolve, or would some refuse to accept the technology forever?