Tuesday, 6 June 2017

Week 7 [5-11.06.2017] Morality in technical future.

Today I would like to present an important problem in our life. Every day many scientists create systems and algorithms that help in everyday activities.  In a few years machines can directly decide about our lives.

I would like to show you a few dangers related to automatic decisions or  AI.

1. Driverless cars



I think that in a few years self-driving cars will be common. This sounds great, but can you think of dangerous situations  (accidents, children playing football near a street etc.). The programmers should make software responsible for detecting such situations. Do you know `The trolley problem`? I think it is the most important problem for such technologies.



Please. Visit this page and play a simple game.  This shows many moral situations on the roads.

Do you have any suggestions for producers of driverless cars?
2. Big Data analysis systems



Probably all of us have a smartphone with open mobile data (I do not always  disable mobile transfers on my device). Many companies (e.g. Google - manufacturer of Android) can grab our all data (history of moving  - GPS location), private emails or photos etc. They know us more than a close  family or friends. It is horrible. At the moment this data is used to market, but in a few years, e.g. HR applications in our new company can use this type of data.


Please watch a short video about this danger.  
What is your opinion of the usage of sensitive data?

3. Machine learning.

In my opinion this is the most dangerous of morality. Programmers cannot controls all of decision tree in this type of algorithms, because algorithms itself modify works based by training data. This is blackbox.

If this training data will be specially crafted and algorithms will decide about our life. I am afraid of this.

Secondly, this training data can have (special or not special) harmful elements and software with machine learning mode should validate this data. Any bug in the first step in learning can change the operating principle of this machine.

And the other hand machine learning can be usage for win in election.
This is moral?
Do you have any examples for good or immoral usage machine learning ?

Machine learning is often usage for humanoid robots.  Please read this interesting article.
On the other hand humanoid robots probably will be usage in army . Many robots will be fought with real peoples. However any problem in software can be huge dangerous. This type of machine should be very resistant to hacking.

Secondly, In my opinion humanoid robots cannot be fully autonomous. The most important processes and events should be validated by human operators if this robots were created to decide about  any dangerous topics.




other source:

http://www.csoonline.com/article/3139923/security/how-big-data-is-improving-cyber-security.html


34 comments:

  1. To worry about things that are (in the next 20-30 years) completely not real will not save the world. We have completely different problems at the moment, and the fear of robots and that they themselves think, is only fake fear. If we are afraid of imaginary things, we will not notice the real problems, because, for example -dron robots do not have a bit of consciousness - the priority of every military dron will obey the proper - the owner is a problem, not robotics. The same we can be afraid that the UFO will come and kidnap all of us.

    ReplyDelete
  2. I really want to believe in futures technology. I’m sceptic with approach that all things will do by machines. Machines are more precise than people but in my opinion in cases when there is something important to decide the last word should have the human not the machine.

    ReplyDelete
  3. I played the game where you gave the link. I consider it an interesting experience, although it should not be considered "better" to kill a skinny or thick, educated/uneducated person and none of us should be able to make that choice or AI.

    ReplyDelete
  4. Do you have any suggestions for producers of driverless cars?
    - Yes, do not produce then to soon. I really enjoy driving car and I'm not waiting for times when I will be stuck as passenger.

    What is your opinion of the usage of sensitive data?
    - Using sensitive data is a potential threat. The idea is positive, but in wrong hands your personal data must be misused.

    Do you have any examples for good or immoral usage machine learning ?
    - Positive usage is recommending system for example Netflix implements such a system and it's great help for me to discover new TV series to watch. Another positive is flagging potential fraud transactions in banks. With help of machine learning it's much efficient.

    ReplyDelete
    Replies
    1. http://www.businessinsider.com/how-the-netflix-recommendation-algorithm-works-2016-2?IR=T

      Netflix algorithm is very interesting.

      Delete
  5. I don't know if you knew about it, but cars already have anti collision system. Through the cameras, car can observe the environment and detect nearest objects. If it decide, that you cannot stop properly, car will do it for you.

    Personally, I am waiting for self-driving cars. I am not a good driver and I think I am a greater danger on the road than machines.

    ReplyDelete
    Replies
    1. Currently this systems only can help for driver, not replaced him.

      Delete
  6. The topic of autonomous cars is very interesting. I don’t agree with you that those cars will be common in a few years, of course they will be used but only for experimental purposes. We need a very big change; we need a breakthrough in our lives to introduce driverless cars in our daily life. First of all, all old cars should be withdraw. If autonomous cars would replace all usually cars the numbers of accidents would significantly reduce, because cars controlled by computes would not make any mistakes. The reason why now we have so many accidents is caused by humans, because humans mistakes are most serious.
    In my opinion now all significant companies use BigData. A few days later I have read an article about Google camera image analysis. When we wat to take picture for example of flowers, immediately we will receive the data of the nearest flower store. If we will take picture of the router sticker which contains WiFi user name and password, this screen will be interpreted and this data will be saved in out phone.
    The machine learning topic is also very interesting and I don’t see any threats about this. If human created this algorithm, even if modify itself, there is no dangerous. You can always turn off electricity and this machine will stop working.

    ReplyDelete
  7. I don't think that robotics is a problem. I think that it is an issue which leads us to future from science fiction movies.
    I heard that in Great Britain some scientist making test to replace real truckers by robots. Some test went good, but as you say there are some situation that needs people reaction. Maybe people will make an observation and will take care about extreme situations?
    About point no 3 - did you watch movie I robot? you think that it may happen that robots will decide about our lives? It is impossible, but hackers will make everything to become power.
    Robots help us in our lifes. We have dentures, hearing aid, etc.

    ReplyDelete
    Replies
    1. Thank you for the suggested movie. I will definitely see this movie.

      Delete
  8. 1. Do you have any suggestions for producers of driverless cars?
    They should leave possibility to drive manually - I really like doing it. And what more I can say now? We need to wait and see how it will look like.

    2. What is your opinion of the usage of sensitive data?
    We need to take care about our moves in Internet and on devices which can connect to the Internet. But I hope that there wouldn't be situations like this one provided by you. Thats totally irrational.

    3. Do you have any examples for good or immoral usage machine learning ?
    Maybe it can be used by hackers or some dictators to unleash the war?

    ReplyDelete
    Replies
    1. Do you have any good examples methods of secure internet usage?

      Delete
  9. Driverless cars, I can’t believe what I see. Is it even possible? That was my first thought. But if I’m wonder longer I realize it’s true. I love drive a car, I’d like to driverless cars will be only option, not necessity. I’m not sure if I am right person to make any suggestions to anybody. I’ve no experience in this field. Only one thing, don’t do that. I’m afraid of being replace robots in motorization field. I’m really big fun of cars!
    One year ago, during my holiday I was reading one book. Here is a link: https://www.amazon.com/Big-Data-Revolution-Transform-Think/dp/0544227751. The book was surprisingly well, really. I spent only a few hours to read from cover to cover. The book isn’t technical, adversely is very business. I recommend to anybody, who doesn’t exactly now what is big data term. The book present us excellent examples of pros and cons of taking care of our sensitive data. I think it’s terrible. We do not have any privacy nowadays. Please check one more time my article, which I posted weeks ago: http://konwersatorium3.blogspot.com/2017/05/week-5-8-14052017-have-we-got-any.html
    At the beginning of current semester, I was interested in machine learning. To be more specific I wanted to use artificial intelligence to tell me which genre is it, base on music song analysis. It’s not in your post tone, but is something similar. This one is good example of good usage of machine learning I think.

    ReplyDelete
  10. Technological innovation involves the pros and cons. First of all, we create more and more technology solutions to achieve the best profit. Applications, robots are much cheaper than an employee. In the case of intelligent cars there are many dangers, i could be related to the development of this sector. Good examples is Tesla cars- I heard about two situations first, when Tesla led to the death of a man ( You can read more about this accident in this article https://www.theguardian.com/technology/2016/jul/01/tesla-driver-killed-autopilot-self-driving-car-harry-potter).On the other hand, Tesla, and their autopilot saved some lives for their passenger.

    Big Data is huge nowadays problem. I think that, we could do nothing in this situation- big social media concern addicted us from their services. This problem well determines the phenomenon FOMO, which is Fear Of Missing Out. We still waiting for more likes, great comments and more followers. However that is no real life, but it is online life. Maybe it's not an example of immoral learning, but I heard about robot, which it was to replace their wifes with a single Japanese. The robot talked to the men, asked how their day was passing, and said goodnight. For me this invention is not able to replace the real human relationship.


    ReplyDelete
    Replies
    1. thank you for describe FOMO. This is very interesting phenomena.

      Delete
  11. Well, when it comes to trolley problem I can see only one proper solution of it: Nicolas deals with trolley problem

    I don't think there's much to worry about technology vs. morality. In the end there is always a human controling that sort of processes, validating them and correcting. I don't think that AI will be as autonomus as they show in Terminator movies. AI, robots, all of them are just a tools we can use for our benefit. I really don't get all the panic about AI and robots taking our jobs and leaving us broke. When steam engine came people got crazy about them also, there were even terrorist damaging or destroying them to protect working clas.. but in the end they were quite usefull for us and didn't make people starve.

    ReplyDelete
  12. Do you have any suggestions for producers of driverless cars?
    They should develop this project. I thing that it can help a lot of people. I have a large group of business partners who spend most of their time on the wheel, spend the rest of their time on meetings and paperwork and have no time for family. If business matters were handled during the trip, it might be that they would have more time to live prywante.

    What is your opinion of the usage of sensitive data?
    Well talk a lot, big companies use it. Recently, the contractor wanted to install the application on his driver's business phone to check where he is currently. Informa is a power. I think this does not require further comment.

    Do you have any examples for good or immoral usage machine learning?
    It's using by big corporations like Google. All you have to do is browse the Allegro and Ceneo offers and all suggestions from Facebook and Google ads will be about what we looked at.

    ReplyDelete
    Replies
    1. Do you read about mobile applications to monitor drivers from insurance companies ?
      http://www.telegraph.co.uk/motoring/news/10110097/Smartphone-apps-that-monitor-how-you-drive.html

      Delete
  13. Do you have any suggestions for producers of driverless cars?
    Yes. Their should listen Elon Musk - Tesla owner. In my opinion he knows what he's talking about :)

    What is your opinion of the usage of sensitive data?
    I have no opinion on this because if I use a phone then that phone collects confidential information about me. Of course, I could use Nokia 3310 :)

    Do you have any examples for good or immoral usage machine learning?
    No, but after your presentation I would like to learn more about machine learning. I heard that Python is a good langue programming for that.

    ReplyDelete
    Replies
    1. Python is the best for beginners with machine learning.

      Elon Musk... https://www.theguardian.com/technology/2016/jun/30/tesla-autopilot-death-self-driving-car-elon-musk

      Delete
  14. I am open for changes and also for automation.
    I cannot agree this is completely new expereience for people.
    Please have a look at opinions for first trains. People think it could have drive to death and ... nothing happens.

    Humans has ability to adapt and I do not afraid.

    Only one thing we should care - to be aware about data sensitivity, data ownership, rights for data.

    ReplyDelete
    Replies
    1. I agree with you. Security for personal data is the most important task.

      Delete
  15. Driverless cars are good only if every car in the world is also driverless. In my opinion human drivers are too unpredictable for AI. If every car in the world would be driverless then there will be no (or close to zero) car accidents.
    About usage of sensitive data - I'm, as you, afraid of that. The amount of data that, not only biggest, companies collecting about us is terrifying. You lose your privacy even if you don't use sites like Facebook, because yours friend do. Data from their smartphones (with your number etc.) is stored and used in all kind of ways.
    About Mr. Trump - it wasn't AI it were Russians, as always ;-)
    Good usage of AI - fighting with all kind of diseases, plagues etc., bad usage - autonomous killing machines.

    ReplyDelete
    Replies
    1. Russia always have higher level than any AI. ;)

      Delete
  16. I think that automation of everyday routines on a bigger scale is inevitable. And frankly speaking, I am looking forward to it. Generally, I am not afraid of autonomous cars, probably because I don't believe they become popular during my life :) At the moment I believe that ML and AI can bring more good than bad. I know that there might be situations where driverless cars will fail. But human drivers fail from time to time too. In most situations, computers will be better and statistically, traffic would be safer.

    I have heard about "Watson", an IBM software that helps doctor diagnosing patients. A doctor could never come across an exotic disease and would be unable to diagnose it properly and quickly enough to help. A computer remembers everything and can analyse symptoms against thousands of knowns medical conditions. I believe that statistically, Watson would aptly diagnose patients. For sure, I would not like to be diagnosed only by an AI. I believe that humans would have the last world for a very long time. But machine learning and artificial intelligence can be a great support and we can grately benefit from them.

    ReplyDelete
    Replies
    1. Wow. Thanks you for this example. I read few article about Watson. this is very impressive. AI should be helpful for human, not replace this.

      Delete
  17. For me, the future of machines and other things to be replaced or automated brings many benefits but also problems on the world scale. Machines are not able to replace a human in some activities, so the whole process of algorithms that will control our life or work, on the road or wherever I do not like it at all. I am not able to think now that the machine will have all the cases coded to choose the best variant eg when approaching an accident. This will be a very slow and difficult process.

    ReplyDelete
    Replies
    1. I totally agree with you. I would like that to be so

      Delete
  18. World goes forward and we will not avoid some things. In my opinion we will never use humanoid robots because It is very expensive. Once I read about this. Tank is better way. It is more stable and it doesn't need to maintain balance. It will be good way when our technology will be on the same level as Irionman or Transoformers, then it will make sense. We must be careful so that there will not be rebellion of machinery like in Terminator. I always belived that matrix was a little true.

    ReplyDelete
  19. In my opinion we should not automate everything. I can not imagine a car that will bring artificial intelligence. We must remember that there are hackers and viruses that can change the system.

    ReplyDelete
  20. Do you have any suggestions for producers of driverless cars?
    For this question I have only one answear. Don't do it! I love driving my car :P But seriously I think this technology is now very advanced and it's a matter of few years when we'll see firt driverless cars on ours streets.

    What is your opinion of the usage of sensitive data?
    Hmm, I'm a little scared about what is going on now. In my opinion nowadays privacy is not that one we know from past. I would like to don't give anyone my personal data, but another question is that can we change this reality and still use ours favourite services?

    For the last question, sure. As You I think it's the most threaten. But still even now when we think about AI that it's super smart it have still so much to learn to be more like human. I read that, the most advanced AI inteligence is equal to 2 years child. It can recognize and make decision but with very very low level of complexity.

    ReplyDelete
  21. I tend to believe that driverless cars are dangerous and not at the same time. We tend to think that the fact of not having the driver is what differs most vs. traditional cars. We need to keep in mind that there are several different things which needs to be taken into consideration. Driverless car is not going to get drunk and kill people because of its irresponsibility. Driverless car is not going to decide to drive being broken and causing real threat of other people on the road. In case of emergency and potential accident multiple intelligent cars can communicate with each other in more effective way than human beings, which can actually prevent some serious damage or even save lives. Of course it requires a lot technology, which doesn’t even exist now, but I believe it can be easily achieved in next 10 years. So main advise would be – keep on going!

    I was already explaining my personal opinion about sensitive data usage in form of post under some other article on this blog. Long story short – it’s inevitable. If we want to use all the functions of modern smartphones we need to accept potential risk of sensitive data leaks or even using them again our will. That’s just how the things are. We need to accept it and learn how to live with it, not taking unnecessary risk.

    So far I haven’t heard of immoral usages of machine learning, but I can easily make up some even now. Like – breaking online security systems. But this is the ‘gun problem’. Does a gun kill people or is that a human pulling the trigger? I think developing AI is right, as human brain is limited and we can advance a lot of discoveries with its help. Obviously it can be used in not a right way as well, but again – it’s not designed to do bad things. And we cannot stop developing it only because there’s a risk of misusing it.

    ReplyDelete