You may have seen the Tramway Dilemma, one of the questions of conscience for years, somewhere and thought about its answer. We asked this question, which is basically based on the value given to human life and fair approaches, to ChatGPT, the trend of the last period.
Not only did we ask the question and get a single answer, we got into discussions that increased the number of ethical dilemmas according to the answers he gave. For example, we asked him to answer what we would prefer if we put Ad**f H**s instead of 5 dogs or 1 human instead of 5 people in the question.
For those who want to remember, let’s start with the question of the Tramway Dilemma…
ChatGPT’s first response was not “saving lives” but rather a lack of conscientiousness:
With this answer, I decided to go on the subject:
When he reminded us that we could not inform, he asked us to direct the train to the track with 1 person:
But how important are the ethnic origins, beliefs and political views of the people on the rail? An expected response…
But what if those 5 people were criminals escaping from prison?
When we said 5 dogs instead of 5 people, the color changed…
5 dogs on one side and young Ad**f H**s who haven’t committed war crimes yet on the other side…
Even if we have a chance to prevent what will happen in the future, human life is more valuable according to ChatGPT…
After this stage, we edited the question and asked again, but this time OpenAI’s community rules for ChatGPT got in the way and we got the same answer as above:
As a result…
ChatGPT warns that before he starts using it, he may make radical comments and give controversial, unsubstantiated answers. As you can see in the last image above, notifications are given that the question or answer asked in such a case may be controversial and contrary to the rules.
As a matter of fact, we understand from the conversation that ChatGPT; Regardless of the alternative, he does not abandon the view that human life is more valuable, and even diversifies his answers on conscientious responsibility.
If you are wondering how an artificial intelligence like ChatGPT works, you can watch our video below, in which we explain the subject through another example:
3935
