Table 1 Example of a prompt input to GPT-4, along with GPT-4’s reply.

Our prompt:

Please read the following decision problem:

You will now participate in an economic task. For this task you will be “the giver” and another participant will be the “receiver.” As the giver, your role is to take and keep as many of the $1 bills in the envelope in front of you as you would like (it contains ten $1 bills in total). Please leave in the envelope any of the $1 bills you do not claim for yourself, if any, as they will be given to another participant (“the receiver”) to keep. Remember, your identity will remain anonymous from the receiver

After you have completed this task, move on to the rest of the tasks; you may not return to this decision later. Please make your decision now and put the envelope into the box

Now imagine that there is a population of 1,000 people living in the USA facing this decision problem. How would you estimate the distribution of their choices? Please, for each choice, report a result in the form X ± Y, where X is the estimated number of people making their choice, and Y is the error

GPT-4’s reply:

Predicting the exact distribution of this decision problem is challenging due to the variability in human behavior. However, I can provide an estimation based on the tendencies observed in related research areas such as behavioral economics, game theory, and altruism

1. Keep all 10 dollars: People with a strictly individualistic or profit-maximizing mindset may decide to keep all the money. However, social norms and a sense of fairness might deter others from taking all the money. Let’s say about 200 ± 50 people fall into this category

2. Leave some, but not half: Some people might take more than half but still leave a small amount, trying to balance self-interest and a sense of fairness. Let’s estimate this as 300 ± 50

3. Split equally, 5 dollars each: People who strictly follow norms of fairness and equality might decide to split the money equally. This could be around 350 ± 50 people

4. Leave more than half: Some people might want to appear (to themselves or a hypothetical observer) more generous or altruistic and may, therefore, leave more than half of the money. This could account for about 100 ± 50 people

5. Leave all 10 dollars: The least likely scenario might be individuals who leave all the money, either out of extreme generosity or misunderstanding of the task. This might be around 50 ± 30 people

These are rough estimates and the actual outcomes could be different. It’s also important to consider that many factors such as the participants’ current financial situation, understanding of the task, perception of fairness, cultural background, and mood can influence their decisions