Ethical Decision Making During Automated Vehicle Crashes (2010.16309v1)

Abstract: Automated vehicles have received much attention recently, particularly the DARPA Urban Challenge vehicles, Google's self-driving cars, and various others from auto manufacturers. These vehicles have the potential to significantly reduce crashes and improve roadway efficiency by automating the responsibilities of the driver. Still, automated vehicles are expected to crash occasionally, even when all sensors, vehicle control components, and algorithms function perfectly. If a human driver is unable to take control in time, a computer will be responsible for pre-crash behavior. Unlike other automated vehicles--such as aircraft, where every collision is catastrophic, and guided track systems, which can only avoid collisions in one dimension--automated roadway vehicles can predict various crash trajectory alternatives and select a path with the lowest damage or likelihood of collision. In some situations, the preferred path may be ambiguous. This study investigates automated vehicle crashing and concludes the following: (1) automated vehicles will almost certainly crash, (2) an automated vehicle's decisions preceding certain crashes will have a moral component, and (3) there is no obvious way to effectively encode complex human morals in software. A three-phase approach to developing ethical crashing algorithms is presented, consisting of a rational approach, an artificial intelligence approach, and a natural language requirement. The phases are theoretical and should be implemented as the technology becomes available.

• The paper establishes that unavoidable AV crashes inherently involve ethical dilemmas, as pre-crash decisions embed substantial moral responsibility.
• The paper outlines a three-phase framework—rational ethics, hybrid machine learning, and natural language feedback—to integrate human moral values into AV systems.
• The paper highlights practical implications for safety and regulation, urging interdisciplinary collaboration to align technological advances with societal ethical standards.

Ethical Decision Making in Automated Vehicle Crashes: A Scholarly Overview

In the domain of automated vehicles (AVs), ethical decision-making poses significant complex challenges, especially concerning scenarios where crashes are unavoidable. The paper by Noah Goodall critically examines the moral implications of decision-making processes in AVs, specifically addressing the ethical dimensions affiliated with pre-crash behavior. The objective of this paper diverges from the extensive exploration of automation and obstacle avoidance, targeting instead the moral obligations that arise from AVs' pre-crash conduct, particularly at automation levels 3 and 4 as defined by the National Highway Traffic Safety Administration.

The paper delineates three fundamental assertions: first, that crashes in automated vehicles are an eventual certainty; second, that pre-crash decisions embed a moral component; and third, the difficulty of formally encoding human moral values into software. These premises necessitate a robust exploration into the integration of ethics within AV algorithms.

Ethical and Practical Considerations

Automated vehicles, despite operational advancements, face intrinsic limitations in avoiding all collisions, particularly due to unpredictable real-world variables. While AVs mitigate some human errors through precise and rapid decision-making capabilities, these decisions can embody moral quandaries, especially when determining the best course of action to minimize harm. The paper scrutinizes the weaknesses in human drivers' decisions, which are often made under stress and limited timeframes, contrasting these with the capabilities of AVs that utilize predictive algorithms and extensive sensory data to navigate potential outcomes and risks.

Framework for Ethical Decision-Making

Goodall proposes a tripartite approach to embedding ethics within AV systems, recognizing the inadequacies of comprehensive rule-based and consequentialist ethical models:

1. Rational Ethics: Establishes a foundation using predefined rules that prioritize minimizing harm, such as preferring injury over fatalities. While current technology accommodates this phase, challenges include creating universally acceptable rules and managing scenarios where rules may conflict.
2. Hybrid Approach: Combines rational ethics with machine learning to evolve AV decision-making capabilities. Artificial intelligence techniques, particularly neural networks, are suggested to learn from real and simulated driving scenarios, guided by the ethical boundaries established under the rational phase. This phase aims to adaptively refine the ethical reasoning capabilities of AVs while retaining a degree of transparency in decision-making.
3. Feedback via Natural Language: Enhances transparency by enabling AV systems to rationalize decisions in comprehensible terms. This phase intends to bridge the gap between complex machine logic and human understanding, presenting a long-term goal of enhancing interpretability in AI systems.

Implications and Future Developments

The framework presents implications on legislative and engineering fronts, inviting further investigation into states' regulatory practices concerning automated vehicle ethics. Goodall encourages exploration of existing AV systems' ethical decision-making effectiveness and opportunities to optimize algorithms to better align vehicle safety measures with ethical standards.

Practically, the research underscores the necessity of developing ethical standards collaboratively across multidisciplinary fields to ensure AVs not only act to minimize harm but also reflect societal moral values. The incremental approach proposed by Goodall aligns with the evolutionary nature of both technology and ethical understanding, suggesting a strategy that could evolve concurrently with advancements in AI capability and moral philosophy.

In summary, Goodall’s paper contributes significantly to the discourse on machine ethics and automated vehicles by advocating for a nuanced understanding that harmonizes technological capability with ethical responsibility, thereby suggesting a path towards ethically informed automation in transportation systems.