Automotive AI Engineers at the Crossroads of Technology and Ethics
As we transition into the age of autonomous vehicles, a pressing question arises: not whether an AI brain can make critical decisions instantaneously, but rather, who will define AI’s sense of right and wrong? The challenge of programming autonomous vehicles for ethical decision-making presents a modern dilemma that engineers must confront.
The Ethical Dilemma
In the realm of automotive AI, unavoidable dilemma situations cannot be ignored. According to experts, a universal moral code for machine ethics and self-driving cars does not currently exist. As Wolf Schäfer, professor emeritus at Stony Brook University, points out, these issues are no longer confined to the engineering domain; they have become critical societal questions.
While ethical theories such as libertarianism, utilitarianism, and Kantianism offer frameworks for understanding morality, the algorithmic implementation of these theories in autonomous vehicles appears arbitrary. Variations in moral preferences across different cultural contexts further complicate the design of a universally acceptable vehicle control system.
Rethinking Engineering Education
Schäfer emphasizes the necessity for change in engineering education, particularly in areas related to AI design and ethical considerations. The shift to autonomous vehicles is projected to take over a decade, during which time it is crucial to plan for the expanding AI landscape. In the United States, approximately 40,000 fatal motor vehicle crashes occur annually, significantly outpacing deaths from homicide, plane crashes, and natural disasters combined.
To address these pressing issues, Schäfer’s Automotive Ethics VIP (Vertically Integrated Projects) team has been integrating moral philosophy into engineering curricula since 2020. The team employs model cars equipped with sensors and cameras to simulate real-world driving scenarios, including moral dilemmas, thereby allowing students to develop algorithms that guide the vehicles’ decisions.
Defining Moral Machines
Schäfer categorizes machines into three distinct types: immoral, moral, and rightful. Immoral machines act against ethical norms, while moral machines adhere to certain ethical guidelines. Rightful machines, on the other hand, are those certified by society to operate on public roads.
These distinctions necessitate discussions about ethical questions, a current gap in engineering education. Unlike universally accepted scientific laws, ethical rules vary significantly among cultures, complicating the establishment of a consensus.
Collaboration Across Disciplines
Schäfer advocates for a multidisciplinary approach, where engineering students collaborate with scholars in the humanities and social sciences. This integrated model not only enriches the engineering curriculum but also enhances the educational experience for students in the humanities.
Students involved in the Automotive Ethics Lab express a passion for integrating AI into daily life responsibly. For instance, one student, who owns a Tesla, is actively developing a machine learning model designed to serve as an ethical decision-making component within their simulation.
The Road Ahead
As the Department of Technology evolves to include a focus on AI and society, the need for ethical deliberation in AI decision-making becomes increasingly crucial. Schäfer highlights that current automotive technology, where vehicles operate as computers on wheels, requires a nuanced understanding of moral choices in dilemma situations.
Despite centuries of philosophical inquiry, a universal moral theory remains elusive. The challenge lies in translating philosophical theories into algorithms that autonomous vehicles can use effectively. As the field of engineering continues to advance, it is imperative that ethical considerations are not left solely to engineers, but rather integrated into the broader educational context.
