This is the first time I have seen a man in technology speak honestly about this, there are issues from replacing real jobs, to the danger posed but machines, thank you Devin.
This is such a big conversation, as I know many people that I would never want to make moral decisions on others. And that makes me wonder that of the right people are involved, maybe the rules about morality can be considered better.
I think the important distinction is some value systems can be learned and some constraints are hard coded. What to teach and what to hard code? What he is discussing is the most difficult issue with artificial intelligence.
Morality is forever changing. Our perspectives, priorities, everything - always changing. So I can see how an algorithm could show a robot or AI or whatever between our standard right and wrong, but I dare say everything else gets a bit murky. Humans aren't even great at figuring it out!
Humans get better too, but not as quickly. There is a great book, Thinking fast and slow that covers many of the fallacies you mention. It will be interesting to see which ones make there way into the computers we train.
all you have to do is look at movies and other forms of media to see that our morality has been chipped away at for a very long time... still, the idea of genetic algorithms sounds very interesting.
Users should be given choices so that the code will run according to their own moral beliefs/preferences. That should be specified before using the program by the users in some kind of settings.
Amazing, thank you. You made me think about Battlestar Galactica newer series again, when the cylons were the topic. Because in the series years ago people made them (organic androids) in theirs image. Like god created people in his image. But... Humans - what if we are AI created in omage of our creators? To create humanity? Kind of subjective morality? We will never discover that and thats amazing :)
sensor improvement needed with less windows like using more solid than magnetic unless magnetic with properly shielded. electronic shielding of processor and take sound out of sensor and keep bimetal sensor in sensor more for example as guide in sensor technology but when automate weapon system then no stopping possible
I would rather have a philosopher describe morality than a computer nerd. I think he's missing an important point. Morality is the weighing of various values against each other in the situation at hand. The hard part is finding universal values. The values for humanity may not be the values for AI, or the values for the life that AI is programmed to honor. After all humans will be temporary in the scheme of life. AI may find it morally necessary to severely restrict humanity to protect the life we all are supposed to value. We, the remaining humans, will be objects for scientific study. Why DID they do that?
i think with colission tech could save more life but it is not yet perect but better than without. however the war time come it could quicken the defeat or victory or number of war death and so we must chose just like we did when we inventd gun powder
This is the first time I have seen a man in technology speak honestly about this, there are issues from replacing real jobs, to the danger posed but machines, thank you Devin.
Thanks Kirsten
This is such a big conversation, as I know many people that I would never want to make moral decisions on others. And that makes me wonder that of the right people are involved, maybe the rules about morality can be considered better.
great shout out Frank, and really important subject
@@bridgettemsanders thanks
@@franklinpgarner thank you..
This is making me rethink all the elements of AI, how can we give these choices to technology!
agree 100%
@@itsmesamrendall I know its seriously flawed
Great talk, great topic. Glad smart people are thinking about these critical issues.
I think the important distinction is some value systems can be learned and some constraints are hard coded. What to teach and what to hard code? What he is discussing is the most difficult issue with artificial intelligence.
Morality is forever changing. Our perspectives, priorities, everything - always changing. So I can see how an algorithm could show a robot or AI or whatever between our standard right and wrong, but I dare say everything else gets a bit murky. Humans aren't even great at figuring it out!
Machines can get better over time, and humans have fallicies - but they have to learn from humans, and we are corrupt!
Humans get better too, but not as quickly. There is a great book, Thinking fast and slow that covers many of the fallacies you mention. It will be interesting to see which ones make there way into the computers we train.
@@devingonier4014 um have you seen America recently, you really think humans get better?
all you have to do is look at movies and other forms of media to see that our morality has been chipped away at for a very long time... still, the idea of genetic algorithms sounds very interesting.
Genetic algorithms are interesting and in certain domains are making a comeback!
@@devingonier4014 yes true, well presented talk.
Users should be given choices so that the code will run according to their own moral beliefs/preferences. That should be specified before using the program by the users in some kind of settings.
Amazing, thank you. You made me think about Battlestar Galactica newer series again, when the cylons were the topic. Because in the series years ago people made them (organic androids) in theirs image. Like god created people in his image. But... Humans - what if we are AI created in omage of our creators? To create humanity? Kind of subjective morality? We will never discover that and thats amazing :)
I love that show too, I think sometimes Star Trek was futuristic too.
Love Battlestar Galactica!
sensor improvement needed with less windows like using more solid than magnetic unless magnetic with properly shielded. electronic shielding of processor and take sound out of sensor and keep bimetal sensor in sensor more for example as guide in sensor technology but when automate weapon system then no stopping possible
Is AI a threat to human morality
I would rather have a philosopher describe morality than a computer nerd. I think he's missing an important point. Morality is the weighing of various values against each other in the situation at hand. The hard part is finding universal values. The values for humanity may not be the values for AI, or the values for the life that AI is programmed to honor. After all humans will be temporary in the scheme of life. AI may find it morally necessary to severely restrict humanity to protect the life we all are supposed to value. We, the remaining humans, will be objects for scientific study. Why DID they do that?
If human beings are sinners, how will they be able to teach machines to be saints?
1. These alleged vehicle deaths’ analogies with morality and AI are seem weird
2. People in hurry are usually tend to be less moral
i think with colission tech could save more life but it is not yet perect but better than without. however the war time come it could quicken the defeat or victory or number of war death and so we must chose just like we did when we inventd gun powder
AI charged with saving humanity, 😂.
When was the last time you saw a scientist look out the window or act like a role model.
Reductive and hurbus.
lol non devil's are Aliens ??