Ethics involves studying and reflecting on how one should act in different situations. We evaluate the actions, the individuals performing them, and the resulting outcomes. When discussing ethics, we often use terms like right and wrong, good and bad. This helps us develop our understanding of what is morally acceptable before we find ourselves in those situations. In this way, we strive to become better individuals and create a better society.
In this section, you will be presented with several ethical questions related to AI and their potential consequences. Choose one or more of these questions and reflect on the challenges they pose.
Case studies: Case 1 – The self-crashing car
You witness an accident where a self-driving car collides with another car, causing significant damage. This raises the question of who should be held responsible for the accident.
Is it the fault of the car manufacturer for not making a sufficiently safe vehicle?
Is it the user of the self-driving car?
Or could the lawmakers be responsible for not establishing adequate laws for both the car manufacturers and users of self-driving cars?
You can test the Moral Machine, which addresses some of these questions in more detailed situations.
Case 2 – The racist AI
The police have started using an AI system to predict criminal behavior based on various factors, including skin color. They discover that the algorithm tends to identify individuals of a specific skin color as potential criminals in 9 out of 10 cases. Before the use of the AI system, only 3 out of 10 individuals from this group were considered potential criminals. What could have happened? Should this be taken seriously?
Case 3 – The neglected patients
You are the Minister of Health, and one day you receive an offer to purchase a watch that monitors the health of all citizens. In the trials conducted, they have discovered and prevented serious illnesses in 90% of the users before they develop into cancer or heart problems. The price is 200 dollars per person, and the company selling the watches has calculated that it can save approximately 500,00 dollars per person who uses the watch.
However, it turns out that the training data is based on caucasian individuals and is therefore better at detecting diseases in caucasian people.
What would you do? Why?
Case 4 – The company
You own a company with 47 employees. Each employee costs you approximately 700,000 kroner per year and works 8 hours a day, 5 days a week. They get sick or are absent from work for various reasons. You receive an offer to purchase AI robots that work 7 days a week, 20 hours a day. These robots cost 1.4 million kroner each to buy. They never get sick and only have 4 hours of downtime each day. What do you choose to do as the owner? Why?
Case 5 – The sentient AI
Tomorrow, researchers discover how to create an AI that thinks and feels at the same level as humans. It has an inner life, and the only thing that sets it apart from humans and animals is that the AI has a robotic body.
Should we treat it like we treat a human or an animal?
Or is it just a object?
Case 6 – The secret choices
Today, private companies are developing various AI models, and they must keep their methods and datasets secret. This means that companies are not required to share information about the datasets they use or how their AI models function. Consequently, we do not know if the AIs are fair and neutral in their use. What dangers does society face when AIs are developed by private companies? What if a country develops its own AI? How should we solve these problems?
Case 7 – The school
Research shows that those from homes with parents of higher income and education use technology better than others. Artificial intelligence will likely amplify these differences. What problems can this cause? Should we do something about it?
Case 8 – Schoolwork
With language models like ChatGPT, students can get pre-written articles, solve math problems, and complete other schoolwork. Why should students continue to do schoolwork?