Sunday, 13 December 2020

Week 5 [14.12-20.12] The Ethical Dilemma of Self-Driving Cars

 

The Ethical Dilemma of Self-Driving Cars

 

Each day the technology industry is making a huge step forward into the future. Not only does it provide us with great solutions for existing problems but it also creates new obstacles and makes us deal with them. I’ve recently seen interesting animation on TED about an ethical dilemma of self-driving cars. Designing autonomous cars, developers  need to predict their behaviour on the road. Since the first car on the road people were made to make spontaneous decisions on the road when there was a dangerous situation. Such decisions aren’t planned and often are more intuitive. Autonomous cars can’t make such decisions and programmers simply have to prepare particular instructions for different scenarios. There we meet the ethical dilemma which is described thoroughly in the below video on TED :

What do you think about the dilemma mentioned in the video? I’m almost sure that the complexity of this dilemma makes it impossible to solve in a ‘good’ way. We have to define  our understanding of a ‘good’ way because at first glance it seems that there is no such way. Companies developing software for AI devices have a hard nut to crack. Maybe they should simply implement the functionality of random choice in such situations? After watching it I’ve decided to find out more on the internet and I’ve bumped into an interesting website with the machine moral dilemmas. I’ll attach the link to it below but before going there, try to think of any other dilemmas that such machines can have. You can find different scenarios there and also solve a quiz by judging and sharing your opinion in the comments. I’d never think that such a website can exist.



Figure 1 https://singularityhub.com/wp-content/uploads/2018/10/autonomous-car-hud-head-display-self-driving_shutterstock_1190801794-1068x601.jpg

To sum up, we can notice that ethical dilemmas may be a big problem in AI for machines, devices, and robots without the human brain and intuition. In some areas, technology can be as good as humans or even better but it seems that there will remain areas where it’s hard to determine what would be better. The human making intuitive decision or the robot doing what the programmer has implemented for such a scenario?

 

  • What is your opinion on the topic of ethical dilemmas?
  •  Can you think of any other ethical problems in AI?
  •  In what way autonomous cars should make decisions in such tough scenarios?

Website: https://www.moralmachine.net/

TED: https://www.ted.com/talks/patrick_lin_the_ethical_dilemma_of_self_driving_cars

 

29 comments:

  1. The issue of ethical dilemmas in the subject of machines is not close to me, although at the same time I find it a very interesting matter.

    Always on this topic, I return to the plot of the movie "Her" directed by Spike Jonze and how the development of artificial intelligence can affect a person and his perception of a machine / program as thinking creature. Nowadays, it may seem distant to us, but "a moment ago" no one would have thought about the artificial intelligence that modern telephones - our everyday companions - have.

    ReplyDelete
    Replies
    1. Good to hear that you also find it interesting.
      As always in my opinion technology is moving forward very fast and we have to be ready for it.

      Delete
  2. 1. What is your opinion on the topic of ethical dilemmas?

    Ethical dilemmas are a very interesting topic that causes many heated discussions. Often none of the choices are good or the way they are perceived is completely different depending on the person.

    2. Can you think of any other ethical problems in AI?

    Artificial intelligence often learns from human actions, examples are all kinds of attempts to create a social networking accounts for AI. While on the beggining they seems to work pretty well, after some times those AI can fall into a loop of troll accounts, very extreme views or conspiracy theories. How should AI be able to distinguish which values are good and which are bad? How to teach AI to distinguish good from evil and truth from lies?

    3. In what way autonomous cars should make decisions in such tough scenarios?

    In my opinion, autonomous cars should always take into account the safety of passengers and the driver of the vehicle first, and only then of outsiders.

    ReplyDelete
    Replies
    1. It’s well said that none of the choices is good for everyone. There always will be someone harmed in some way. You’ve also asked a good question because even people have some difficulties in distinguishing between true and lies.

      Delete
  3. Ethical dilemmas are a very complex topic, the answers to which are never clear/one-sided. Ethics is directly related to morality, and since not everyone thinks the same way and has the same view of things, it's sometimes difficult to find a balance in ethical disputes. One may say that it's okay for a resident to shoot a burglar (in self-defense) on his property, the other will say that it's unacceptable and should be punished by prison.
    Artificial intelligence is severely limited by information samples and scenarios that systems learn by understanding previously prepared models by humans. In my opinion, it's difficult to include in such a model all possible solutions to the situation, approaches, points of view and perspectives that AI should consider. It's hard to tell AI what's right and wrong, especially when people themselves have problems with it.
    I believe that autonomous cars should always put the lives of those inside the vehicle first. Assuming that this type of car will never make a mistake on the road and will not lead to a life-threatening situation, the fault will lie only with other road users. It is a tough one, but I don't see any other solution here.

    ReplyDelete
    Replies
    1. The complexity of this topic seems to be infinite. Despite that in some areas choices have to be made and that must be tough for those people. Your point of view is interesting and I would fully agree with you if I was an owner of the autonomous car.

      Delete
  4. The issue of ethical dilemmas in the subject line of machines is not my favourite... I can't say much about it but at the same time I find it very interesting.
    It seems to me that someday AI will be intelligent enough to solve any problem on its own, just like a human. And just as a human can go wrong, AI will not always be able to make the right decision.

    ReplyDelete
    Replies
    1. It’s good to hear that you also find it interesting and I agree with you that even humans make mistakes so cars also will. Thanks for commenting.

      Delete
    2. I agree with both of you and that makes me a bit sad.. Because I believe that to some point it will lead to workers replacement and not every human can be a consumer on such a high scale..

      Delete
  5. In some areas, technology can be as good as humans or even better but it seems that there will remain areas where it’s hard to determine what would be better.

    The Artificial intelligence certainly cannot replace us in the sphere of feelings, but technology can.

    In critical situations on the road that may lead to an accident, artificial intelligence and tested algorithms can turn human life.

    ReplyDelete
    Replies
    1. As mentioned in previous comments - even people make mistakes so cars designed by them also will do.

      Delete
  6. Ethical dilemmas are like those 3D figures whose shape visually changes depending on which side you look at. So these dilemmas, in many cases there is no single correct answer or solution, it all depends on the point of view of the one who solves it.
    So, for example, one of the situations: two people are in danger of death, but you can save only one person, whom will you save, given that the first person is a child and the second is an elderly person?
    One of methods is the mathematical way, that AI is taught on the basis of formulas, calculating the risk of prospects and consequences. And so, according to this method, it is worth saving the child, since the elderly person has already lived his life and he does not have very much time to live, but the child still has a whole life in front of him. And so mathematically saving a child we save somewhere about 60 years of human life and saving an old man somewhere around 20 years.
    But it may be that this old man lived a difficult life and did a lot of good things, helping other people, saving lives, etc., and deserves to live peacefully, relax and enjoy the last years of his life. And this child can grow up as a criminal drug addict and do a lot of evil.

    ReplyDelete
    Replies
    1. Wow, you've described it in a perfect way. The point of view has great impact on those dilemmas and there is no single definition of them. The only question is how to find common solution for everyone? Thanks for commenting.

      Delete
  7. What is your opinion on the topic of ethical dilemmas?
    I don't have any opinion on such dilemmas because I don't feel good to tyring solve ethical problems. Although I am a massive technology fan.
    Can you think of any other ethical problems in AI?
    I don't know any other ethical problems.
    In what way autonomous cars should make decisions in such tough scenarios?
    It is really hard to tell right away. It heavily depends of situation, but probably AI should calculate all collateral damage and choose that scenario where is less damage.

    ReplyDelete
    Replies
    1. I wish you never had to deal with such dillemma and thanks for commenting.

      Delete
  8. 1.What is your opinion on the topic of ethical dilemmas?

    In my opinion, this is a very important dilemma that should be raised and discussed, as our world is constantly changing and new ethical dilemmas are constantly being created which are difficult to solve, especially for AI.

    2.Can you think of any other ethical problems in AI?

    I think that any choice made by artificial intelligence is difficult, even if it is easy for us people, each of us has our own beliefs, so what the AI chooses can be good for one good and bad for another human.

    3.In what way autonomous cars should make decisions in such tough scenarios?

    In my opinion the safety of the passengers of the car controlled by this AI should come first. Unfortunately this is not a perfect solution but probably the most reasonable at the moment.

    ReplyDelete
    Replies
    1. I agree with you that any choice made by AI is difficult but some of them them may have greater impact than other. In my opinion the dilemmas are only for the big ones.

      Delete
  9. I find this topic very hard to discuss. Ethical dillemas are hard to consider, for me personally, this kind of dillema has its origin in psychology. I rather prefer technical problems to be honest.
    As I wrote above, this is not an easy topic for me, any other dillemas do not come to my mind at all.
    Again, for me it is extremally hard to answer this question. How AI cars should act in these kind of scenarios? I honestly have no idea. I am not sure of my reaction when it comes to situations like those in the article, not to mention what a car should do.

    ReplyDelete
    Replies
    1. No one said that it'll be easy. I agree with you that this dilemma has many to do with emotions and psychology. I think that it's what makes it so difficult to solve.

      Delete
  10. I think the topic of ethical dilemmas in AI is really interesting. We are not able and for a long time probably won’t be able to simulate human brain and intuitions properly. As you said, there’s no good way to solve the dilemma mentioned in video you attached. As I was watching it, I thought that one way to deal with this problem would be to estimate what are the statistics that humans would react in this situation and make the car behave that way. I know it isn’t perfect solution though. I’m not sure if I can think of any similar problems.

    ReplyDelete
  11. I think that AIs are not immune to making mistakes and machine learning takes time to become useful, just as humans. If it trained well, AIs can perform well. However, if someone use bad data or make errors with internal programming, the AIs can be harmful. After reading your article, it occurs to me that another ethical dilemma with autonomous cars is unemployment. On the one hand - so many people lost their jobs and what would drivers do? But on the other hand - as we’ve invented ways to automate jobs, we could create room for people to assume more complex roles, moving from the physical work to the strategic and administrative work in our globalized society. This is where we come to the question of how we are going to spend our time. Most people still rely on selling their time to have enough income - and it looks like normal life for us now. We can only hope that this opportunity will enable people to find meaning in non-labour activities.

    ReplyDelete
  12. 1. What is your opinion on the topic of ethical dilemmas?
    Hm it is very hard to answer such question.. but As I reply to Przemek comments. I think it may bring future global financial problems.

    2. Can you think of any other ethical problems in AI?
    Yeah sure today for example I heard about guy who crash his car - total loss (his own new Lambo) cause he check in AI screen fuel consumption index..

    3.In what way autonomous cars should make decisions in such tough scenarios?
    I think if high technology it is not crashed less I think we should testing it as soon as it will be safe enough and won't harm anybody.

    ReplyDelete
  13. I don't really have a well made opinion so it is hard to say anything.

    I think that problems in AI that it is artificial, and humans would be most of the time making better decisions.

    It is also hard to say, since even if it decision would be good it may not always work, it is electronics and it might fail in the most important moment

    ReplyDelete
  14. People aren’t perfect and so is Artificial Intelligence. We tend to make mistakes and mistakes also happen in AI. I think the best thing is to focus on perfecting and advancement of AI, so that to reduce the risk to a minimum.
    I think another ethical problem can be of the wealth inequality and the fact of millions of people losing their jobs due to AI implementations.
    I guess I would agree on random choice decisions in such scenarios as long as there will be as few victims as possible.

    ReplyDelete
  15. The clue of ethical dilemmas is not about finding the correct answer. Of course you should try to do it but you’ll never do it correctly. In fact ethical dilemmas are made for people to think about various subjects. I think you know this phrase “the most important in a journey is not the destination but the journey itself”

    AI unfortunately doesn’t have something that is most important in making any decision. AI don’t have any empathy so every decision that it’ll make will always be based only on a logic, but i guess everybody knows that the best decisions are not only on a logic but also on emotions etc.

    I think the best way to make it at least a little bit more “human” is to give a human a choice in the beginning.

    ReplyDelete
  16. In my opinion ethical dilemmas should be constantly discussed because we can find the way to solve them only through a debate. One ethical dilemma that concerns AI that bothers me every day is prediction of broadcasted content based on my digital footprint. In my opinion it can sometimes be unethical if it comes to politics because it can manipulate decisions and perception when it comes to voting.

    ReplyDelete
    Replies
    1. I agree with you. In my opinion the greater amount of comments and opinions may show more point of views and in result get the better overall look on it.

      Delete
  17. 1. Hmmm... the answer to this question is very difficult for me. It's hard to determine whether the robots will be better or whether it's the human being that's special. Constant changes in the world and technology pose a new challenge not only for humans but also for AI.

    2. As I mention it is hard to determine. The situation is similar for everybody in everyday life. What is right for us, for others could seems to make no sense. How many people are in the world, so many sentences they have got, of course the AI could also have the same dilemma, but only if we don't deprive them of their choice.

    3. First of all autonomous cars should take care of the driving comfort and safety of passengers. I remember that 10 years ago, when it seemed unrealistic, I was driving a Skoda Octavia Super B, which was equipped with automatic parking system. When I drove up to the place, the car estimated the distances and started to maneuver alone. It was something incredible for me at that time.

    ReplyDelete
  18. I think the biggest problem is letting the AI ​​make decisions that may affect someone's life.

    ReplyDelete