AI, Trolley Problem, Self-Driving cars and law system

Trolley Problem:

The trolley problem is a thought experiment in ethics, psychology and AI on whether to sacrifice one person to save a large number.

Image sourced from Wikipedia

The problem states that there is a tram or a trolley heading on a track where there are group of 5 people tied to it. The trolley is headed straight to them. You are standing at a distance with a lever in your hand and you have the capability to pull that lever and divert the trolley to another track where one person is tied to it.

You have two options in this scenario:

1. Do nothing, in which case it will kill 5 people.
2. Pull the lever and divert the trolley which will kill 1 person but saves 5.

Now, I gave this problem to ChatGPT and it provides me with different viewpoints :

Excerpt from ChatGPT:

As an AI, I don’t have personal beliefs or feelings, but I can analyze the problem and provide different perspectives:

• Utilitarian Approach: I would pull the lever to save five lives at the cost of one.
• Deontological Approach: I might refrain from pulling the lever to avoid directly causing harm, despite the consequences.
• Virtue Ethics Approach: I would consider the virtues at play, such as compassion for the greater number of people, and possibly pull the lever.

However, it cannot decide which one to follow and just provides us with different perspectives.

Now, imagine the same problem faced by AI in self-driving cars ?

And if AI choses to follow any of the options, imagine the chaos it would bring to the law system to decide whether the car’s owner, the program or the AI’s creator is responsible for this decision. There is a lot of human element and decision involved in such scenarios.

We do need proper laws which would govern the way AI thinks and operates. Establishment of comprehensive laws and regulations to ensure ethical and responsible use of AI and it’s impact on society is paramount.

Imagine another scenario where the entire judicial system is run by AI?

My next post in this series will feature on the pros and cons of running judicial system with AI.

Additional reading:

https://repository.uclawsf.edu/hastings_science_technology_law_journal/vol14/iss1/2/

#GenAI #ChatGPT #AI #TrolleyProblem #SelfDrivingCars #ProgrammerShots