While scientific and technological advancements in artificial intelligence and robotics are making autonomous vehicles feasible, social and ethical questions are also being raised to ensure responsible development and deployment. In particular, much discussion has centered on how situations of fatal accidents are to be considered with the use of autonomous vehicles in unpredictable traffic environments. In this paper we provide a metaphilosophical analysis of the methodological difference between thought experiments in philosophy and problem-solving in science and engineering in response to whether the infamous trolley problem can be used to guide the ethical of autonomous vehicles for real-life use cases. We argue that such analysis can further mutual understanding and successful cross-disciplinary collaboration between philosophers and computer scientists working on artificial intelligence and robotics.