Developing an Artificial Morality Algorithm for Self-Driving Cars

Overview

With recent developments in self-driving cars, a myriad of positives are apparent, such as their eco-friendliness. Currently, these cars still require human oversight at all times, but the future may bring fully autonomous vehicles to the road with little to no human oversight. In the event of this development, additional programs would need to be in place to dictate the behavior of these vehicles, namely something that would allow them to react to scenarios similarly to a human operator. To achieve this degree of autonomy, vehicles would need to know how humans would react to a given scenario. To do this, a survey was conducted, with two main parts. In the first section, 5 statements were given (e.g. “Passengers’ lives should be favored over pedestrians’ lives”), and participants expressed their degree of agreement or disagreement. This section was used to write the algorithm by assigning numerical values to each response and finding the mean score for each statement. Java was used for writing the algorithm due to its efficiency. In the second section, 10 hypothetical moral dilemmas were posed. Similar to the trolley problem, the scenarios had two outcomes, and participants decided which outcome was the most moral course of action. This section of the survey was used to test the frequency with which the algorithm agreed with participants. Information from each scenario was entered into the algorithm, and it synthesized a moral decision by weighting each variable using the values obtained from the first section of the survey. It was found that the algorithm agreed with the majority of participants in 90% of the posed scenarios. In conclusion, though not perfect, this means that it is possible for an algorithm to mimic human morality, which may be a crucial aspect of vehicle autonomy in the future.

Competition history

• AJAS 2024 Category not listed

Related projects

Closest projects by meaning, across every fair and year in the corpus.

Browse more like this

• All Robotics and Intelligent Machines projects
• AJAS Robotics and Intelligent Machines projects
• Every AJAS 2024 project