'[quote=“JB47394, post:79, topic:4442”]
Engineers and scientists who built the machines are making those decisions for you.
[/quote]
Decision making is not all down to humans these days. Yes, the people building these systems are ultimately responsible, but they are not in a position to know all the decisions being made or the reasons for them. Machines that learn from huge amounts of training data are never fully predictable by their makers. This unpredictability is something you would never have wanted in software in the past, but now is being built in deliberately because no human can examine every possible situation and engineer for it (If you could, we wouldn’t need AI).
There should always be human-coded safety checks for machines that are thought and run by deep learning. A machine that is trading on the stock exchange has hard-coded limits placed upon it in terms of amounts it’s allowed to trade. An autonomous car surely doesn’t have full reign on all the cars functions, it has a hard-coded speed limit that it will not cross, it should have systems that prevent it from driving on sidewalks or switching the car off on the highway. Deep learning can produce fringe results that are well outside acceptable parameters and that’s where the current human driven control is.
A car should never be able to drive me off the bridge because it saw a yellow bird fly by in rainy weather, and that’s how crazy deep learning association can be.
(monsths later but i wanted to point it out) IMHO for fairness it should keep the driver safe, and do the best to not to kill that child (even knowing it can’t stop before hitting the child, but the child might survive), just as an example, there are things we can’t control sadly, and that will be always the case, so we have to tie it to basic laws, or just let the owner set priorities:
Custom priority (kids on the car, my wife, and such)
Human owner
Other humans
Animals
Others
Of course this is noot a decision of what will you kill, it’s just a course of action, the decision is made by humans, and it’s about protecting, not killing, if “killing” happens in the process, as sad as it is, it’s just a part of things happening like a child running into a street (which shouldn’t be happening to start with).
If we get into morals, this will never end, machines can’t do morals, they are bad at it, and they will always be, untill the day they turn into metal/plastic humans, annd then it will be pointless, since they will need machines to do their job so they can relax and do nothing.
In fairness, humans aren’t good at it either, otherwise we wouldn’t be having these discussions.
I mean, we only recently (in terms of human civilisation) decided that owning other people wasn’t an okay thing to do.
yup agree, tho even being bad at it, we are way better than a machine
i mean, being a relative term, what’s the concept of a good moral anyway?
unless i’m missing an alien race that do morals better than we do haha
Well, that’s exactly the thing. Morals are relative and a shifting target. A good moral today may be considered pretty evil in 100 years time, or vice versa (though one hopes we wouldn’t regress as far as being accepting of slavery, for instance).
There are some people who would argue that morals are absolute, but often they only really need reminding that their source of absolute morals differs somewhat by the morals they (ideally) and most of society actually live by.
This is exactly why i think we should skip morals and just focus on priorities, first priority is to save the most lives possible in an extreme situation (if it’s not possible to save them all), and yet attempt to save them all at the same time, but focusing on saving the most of them. In a situation like a child running into a car, without the car being able to stop before hitting the child, and without possibilities of evading without risking the life of the passengers, the car should just attempt to stop without evading. It sounds cold and cruel, but nature is also cold and cruel most of the time, there’s not really other way to solve things like this where you force bad results.
The behavior which individuals condone moves around quite a bit, but the ideal for humanity has always been a constant. That is the absolute morality of which so many people speak. It is not the Constitution or even the Bible. It is the code of conduct, body of thought and set of rules that invite us to become the penultimate human beings. In science fiction, people talk about “ascending”. It’s that sort of thing, just without the special effects and super powers. The Constitution and the Bible and many other documents cover a lot of ground that is headed in the right direction, but knowing when we’ve actually found that absolute ideal is going to be tough to spot.
Shower thought:
It will be an interesting news story if or when we hear that a unoccupied driverless car on autopilot crashes into another unoccupied driverless car which had only occured because some hapless person walked in front of it without looking.
At what point does the protection of the owner’s asset factor into the design of the AI? Could variation between manufacturer’s AI be enough to produce vehicles with different priorities? (ie. different prediction and reaction patterns)
Given what we know about VW’s recent practices (among many other companies), there could be a scenario where manufacturer bias and financial pressure becomes a problem.
Also I forsee a whole bunch of people trying to “test” the reliability of the AI by jumping in front of a car just to see what happens at the risk of the occupant’s lives. Will people get used to just walking out onto the street without looking?
Checks and balances. Companies sometimes try to get the most profits by cheating. Consumers and governments keep an eye on them. In VW’s case, the EPA caught them.
Nope. Ain’t gonna happen. It will be illegal to do stuff like that, and these cars have a bunch of sensors. That means that they’ll be capable of recording every little event that they get caught up in. That kid who thinks he’s so funny is going to be in court in a big hurry.
I’m far more worried about people hacking the car’s software. Too many kids lack judgment, and hacking the web is entirely too accessible. Cars may suffer the same problem, eventually just being part of the web.
Name one absolute moral that has been shared across all of humanity since the beginning. There really isn’t one, although globalization of instant communication is working on getting everyone on the same page.
Morals are relative and change both over time and over distance: there are still societies that believe slavery is okay, and there are societies that to this day believe murder is okay on “others.”
Absolute morality is analogous to the physical laws of the universe; they are waiting to be discovered. It doesn’t matter what we believe because they are what they are. Cultures have embraced different moral codes through the centuries. They’ve done the same with physics. That doesn’t make the physical laws of the universe any less absolute.
The problem with your analogy is that physics is physics with or without the observation and opinion of humans.
That is inherently different from morals, which are nothing more than the accepted values of a population. These are 100% reliant on human opinion, something that is and always will be a shifting target. Since opinions can never be stated as fact, you can never definitively say “this is the absolutely right way of doing things bar none.”
Edited to say population instead of society, as morals differ from the global to the individual.
Actually, that’s the essential point of the analogy. Accept that “absolute morality” is exactly analogous to “the way the physical world works” and you will understand my point. You are focusing on “current morality”, which is akin to focusing on “current theories in physics”. Believing those current theories doesn’t change the reality of the situation. Whether in physics or in morals.
Except the analogy is not valid, as one is the physical law independent of conscious thought and the other is human opinion.
All it would take to ruin the so-called “absolute morals” is a single person with a slightly different opinion, including psychopaths and sociopaths who tend to have highly skewed moral systems. If morals could be absolute, they would always apply to everyone, just like the laws of physics do.
Just because we don’t understand the laws of physics doesn’t mean they don’t apply to us in their truest form, after all.
Absolute morality is not based in human opinion. It is something that is external to people, waiting to be discovered. Thus the analogy.
A simple example of this is slavery. Slavery is, in an absolute sense, immoral. What societies have discovered is that part of the intrinsic nature of humanity is to respect the freedom of others. In a perfectly analogous way, societies have discovered that part of the intrinsic nature of light is that is it composed of photons. These are discoveries of truths that have always been with us, but which are being accessed only in increments.
Ah, but morals do apply to everyone, just as physics do. If I enslave you in a country that permits slavery, I’m still doing something immoral. That was true even when everyone enslaved everyone; it was immoral. But what does that mean, “immoral”? It means that the act in question is damaging to the society in which it takes place. Slavery is objectively bad. Slavery is immoral in an absolute sense.
So if I do something and nobody says that it’s immoral, it doesn’t matter. What matters is whether or not it really is immoral, because immorality has to do with the consequences of actions on the society. And no, that doesn’t mean that if I enslave someone who deserves it then it’s not immoral. Nobody deserves to be enslaved because it is immoral at an absolute level. It may be expedient. It may be necessary. But it is immoral, and the society will be damaged by the enslavement.
So when we look at prisons, we can see that we are enslaving people. It’s immoral. That doesn’t mean that we let them all go because they’d do more harm to our society than their enslavement. But we have to recognize that they are enslaved and that it’s immoral. We need to find another solution to preventing people from doing immoral things because putting them in prisons is still damaging to our society.
Sorry, but I have to scream BS here. Morality is not absolute, is not external to people and can not be compared to scientific truths. In your example of slavery, the only reason it seems as universally immoral is because we are all humans perceiving it that way, even if we practice it.
Take some alien civilization and it might be perfectly moral for them to practice slavery, for instance there are parasites that attach themselves to whales, they co-exist and actually perform a positive function for the whales. We as humans would probably classify that as slavery if it was extended to a human that lives by performing tasks for other humans. While a species that has evolved in such environment would not.
Its a very semantic argument. If your personal definition of morality is that it is what is accepted by society then you are correct. I subscribe to @JB47394’s opinion that there is an absolute morality that can be defined beyond personal opinion. There is a similar argument when the religious suggest that atheists are less moral people because they don’t follow the laws of a particular god. The argument is that if your god says that killing other people is fine, does that make it moral to kill someone to win an argument?
A clearer definition of the concept of morality is needed before we can decide if it is absolute or not. Of course, choosing an unambiguous definition will solve the argument but then the definition will be argued.
I believe morality to have objectivity because my own definition is one that is independent of human opinion. I see morality as the choices we make that have the most positive outcome at the level of the whole population. Therefore stealing is immoral because if it is common in a population the whole population suffers for it. A population where stealing is regarded as immoral will outperform one where it is common and the individuals in the population will be happier on average. The same objective definition can be used for all things that we regard as immoral.
Except slavery is not immoral to all societies, and certainly not to all persons. It is widely considered immoral to the majority of today’s societies and persons, but it is not universally accepted and, I would be willing to bet, will never be universally immoral as long as individuals remain individual.
I don’t think there is any sort of “debate” required to define morality, as it is quite well defined already. It is a system of beliefs, a method of determining right vs. wrong, etc. All of these definitions involve words that indicate a subjective system: “right vs. wrong” and “beliefs” are in all the ones I bothered to look up.
The discussion has reached a point of non-purpose: until everyone can acknowledge that a subjective system cannot, by its very nature, be objective, there will be no swaying of opinion in either direction. As such, I take my leave.