0 Members and 1 Guest are viewing this topic.
Quote from: hamdani yusuf on 14/10/2019 06:43:23A person gets brain damage that makes him unable to feel pain and pleasure, while still capable of doing normal activities. Is he still considered sentient? Does he still has right to be treated as sentient being? Why so?If you've removed all of that from him, there could still be neutral feelings like colour qualia, in which case he would still be sentient. You could thus have a species which is sentient but only has such neutral feelings and they would not care about existing or anything else that happens to them, so they have no need of protection from morality. They might be programmed to struggle to survive when under attack, but in their minds they would be calmly observing everything throughout and would be indifferent to the outcome.
A person gets brain damage that makes him unable to feel pain and pleasure, while still capable of doing normal activities. Is he still considered sentient? Does he still has right to be treated as sentient being? Why so?
In the case of your brain-damaged human though, there are the relatives, friends and other caring people to consider. They will be upset if he is not protected by morality even if he doesn't need that himself.
That's the very problem identified by philosophers critisizing utilitarianism. How can you expect anyone else to agree with your thoughts when your don't clearly define what you mean with sentience, which you claimed to be the core idea of universal morality?
At least you have to define a criterion to determine which agent is more sentient when compared to another agent. It would be better if you can assign a number to represent each agent's sentience, so they can be ranked at once. You can't calculate something that can't be quantified. Until you have a method to quantify sentience of moral agents, your AGI is useless to calculate the best option in a moral problem.
AFAIK, neuroscience has demonstrated that pain, pleasure, sadness, happiness are electrochemical states of nervous systems, and human already have basic understanding of how to manipulate them at will. I think we can be quite confident to say that rocks feel nothing, thus not sentient.
So if the brain-damaged human has no relative or friend that care, e.g. unwanted baby left by the parents, there would be no utilitarian moral reason to save him/her.
I'm not required to spell out what is sentient and in what ways it is sentient. That task is part of the calculation: what are the odds that species A is sentient, and how much does it suffer in cases where it suffers, and how much pleasure does it experience in cases where it enjoys things. AGI will make the best judgements it can about those things and then act on the basis of those numbers. It will look at rocks and determine that there is no known way to affect how any sentience that might be in any rock is feeling, so anything goes when it comes to interactions with rocks.
It's AGI's job to work out those numbers as best as they can be worked out.
Neuroscience has demonstrated nothing of the kind. It merely makes assumptions equivalent to listening to the radio waves coming off a processor and making connections with patterns in that and the (false) claims about sentience being generated by a program.
Do you know how Artificial Intelligence work?
Their creators need to define what their ultimate/terminal goal is.
An advanced version of AI may find instrumental goals beyond the expectation of its creators, but they won't change the ultimate/terminal goal.
I have posted several videos discussing this. You better check them out.
Neuroscience has demonstrated how brain activity would be like when someone is conscious and when someone is not conscious. It can determine if someone is feeling pain or not, pleasure or not. At least it can demonstrate sentience in the standard definition.
Does your calculation include emotional states such as happiness, sadness, love, passion, anger, anxiety, lust, etc.?
You have claimed that the ultimate goal of morality is maximizing X while minimizing Y. But so far you haven't clearly define what they are and their boundary conditions, so it's impossible for anyone else to definitively agree or disagree with you.
QuoteNeuroscience has demonstrated how brain activity would be like when someone is conscious and when someone is not conscious. It can determine if someone is feeling pain or not, pleasure or not. At least it can demonstrate sentience in the standard definition.All it has demonstrated is correlation with something that may or may not be real. If you pull the plug on a machine that's generating false claims about being conscious, the false claims stop. The link between the claims being generated and particular patterns of activity in a processor do not determine that the claimed feelings in the system are real.
I wasn't talking about artificial intelligent machines here. It was experiments on living humans using medical instrumentation such as fMRI and brainwave sensors that can determine when someone is conscious or not, when they are feeling pain or not. We can compare the readings of the instrumentations and the experience of the human subjects to draw a general patterns about what brain conditions constitute consciousness and feelings.
You are talking about biological machines which generate claims about consciousness which may not be true, just as a computer can generate claims about experiencing feelings (including one of awareness) without those claims being true. When you disrupt the functionality of the hardware in some way, whether it's a CPU or a brain, you stop the generation of those claims. You do not get any proof from that that you are narrowing down the place where actual feelings might be being experienced.
Russell & Norvig (2003) group agents into five classes based on their degree of perceived intelligence and capability:1. simple reflex agents2. model-based reflex agents3. goal-based agents4. utility-based agents5. learning agents
I think it's good that you and others are still exploring this. We'll soon be able to put all the different approaches to the test by running them in AGI systems to see how they perform when applied consistently to all thought experiments. Many approaches will be shown to be wrong by their clear failure to account for some scenarios which reveal serious defects. Many others may do a half-decent job in all cases. Some may do the job perfectly. I'm confident that my approach will produce the best performance in all cases despite it being extremely simple because I think I've found the actual logical basis for morality. I think other approaches are guided by a subconscious understanding of this too, but instead of uncovering the method that I found, people tend to create rules at a higher level which fail to account for everything that's covered at the base level, so they end up with partially correct moral systems which fail in some circumstances. Whatever your ideas evolve into, it will be possible to let AGI take your rules and apply them to test them to destruction, so I'm going to stop commenting in this thread in order not to lose any time that's better spent on building the tool that will enable that testing to be done.
In 2012, Oliver Scott Curry was an anthropology lecturer at the University of Oxford. One day, he organized a debate among his students about whether morality was innate or acquired. One side argued passionately that morality was the same everywhere; the other, that morals were different everywhere.“I realized that, obviously, no one really knew, and so decided to find out for myself,” Curry says.Seven years later, Curry, now a senior researcher at Oxford’s Institute for Cognitive and Evolutionary Anthropology, can offer up an answer to the seemingly ginormous question of what morality is and how it does—or doesn’t—vary around the world.Morality, he says, is meant to promote cooperation. “People everywhere face a similar set of social problems, and use a similar set of moral rules to solve them,” he says as lead author of a paper recently published in Current Anthropology. “Everyone everywhere shares a common moral code. All agree that cooperating, promoting the common good, is the right thing to do.”For the study, Curry’s group studied ethnographic accounts of ethics from 60 societies, across over 600 sources. The universal rules of morality are:Help your familyHelp your groupReturn favorsBe braveDefer to superiorsDivide resources fairlyRespect others’ propertyThe authors reviewed seven “well-established” types of cooperation to test the idea that morality evolved to promote cooperation, including family values, or why we allocate resources to family; group loyalty, or why we form groups, conform to local norms, and promote unity and solidarity; social exchange or reciprocity, or why we trust others, return favors, seek revenge, express gratitude, feel guilt, and make up after fights; resolving conflicts through contests which entail “hawkish displays of dominance” such as bravery or “dovish displays of submission,” such as humility or deference; fairness, or how to divide disputed resources equally or compromise; and property rights, that is, not stealing.The team found that these seven cooperative behaviors were considered morally good in 99.9% of cases across cultures. Curry is careful to note that people around the world differ hugely in how they prioritize different cooperative behaviors. But he said the evidence was overwhelming in widespread adherence to those moral values.“I was surprised by how unsurprising it all was,” he says. “I expected there would be lots of ‘be brave,’ ‘don’t steal from others,’ and ‘return favors,’ but I also expected a lot of strange, bizarre moral rules.” They did find the occasional departure from the norm. For example, among the Chuukese, the largest ethnic group in the Federated States of Micronesia, “to steal openly from others is admirable in that it shows a person’s dominance and demonstrates that he is not intimidated by the aggressive powers of others.” That said, researchers who studied the group concluded that the seven universal moral rules still apply to this behavior: “it appears to be a case in which one form of cooperation (respect for property) has been trumped by another (respect for a hawkish trait, although not explicitly bravery),” they wrote.Plenty of studies have looked at some rules of morality in some places, but none have attempted to examined the rules of morality in such a large sample of societies. Indeed, when Curry was trying to get funding, his idea was repeatedly rejected as either too obvious or too impossible to prove.The question of whether morality is universal or relative is an age-old one. In the 17th century, John Locke wrote that if you look around the world, “you could be sure that there is scarce that principle of morality to be named, or rule of virtue to be thought on …. which is not, somewhere or other, slighted and condemned by the general fashion of whole societies of men.”Philosopher David Hume disagreed. He wrote that moral judgments depend on an “internal sense or feeling, which nature has made universal in the whole species,” noting that certain qualities, including “truth, justice, courage, temperance, constancy, dignity of mind . . . friendship, sympathy, mutual attachment, and fidelity” were pretty universal.In a critique of Curry’s paper, Paul Bloom, a professor of psychology and cognitive science at Yale University, says that we are far from consensus on a definition of morality. Is it about fairness and justice, or about “maximizing the welfare of sentient beings?” Is it about delaying gratification for long-term gain, otherwise known as intertemporal choice—or maybe altruism?Bloom also says that the authors of the Current Anthropology study do not sufficiently explain the way we come to moral judgements—that is, the roles that reason, emotions, brain structures, social forces, and development may play in shaping our ideas of morality. While the paper claims that moral judgments are universal because of “collection of instincts, intuitions, inventions, and institutions,” Bloom writes, the authors make “no specific claims about what’s innate, what’s learned, and what arises from personal choice.”So perhaps the seven universal rules may not be the ultimate list. But at a time when it often feels like we don’t have much in common, Curry offers a framework to consider how we might.“Humans are a very tribal species,” Curry says. “We are quick to divide into us and them.”
Philosophy can be perceived as a rather dry, boring subject. Perhaps for that very reason, divulgers have attempted to use stimulating and provocative thought experiments and hypothetical scenarios, in order to arouse students and get them to think about deep problems.Surely one of the most popular thought experiments is the so-called “Trolley Problem”, widely discussed across American colleges as a way to introduce ethics. It actually goes back to an obscure paper written by Philippa Foot in the 1960s. Foot wondered if a surgeon could ethically kill one healthy patient in order to give her organs to five sick patients, and thus save their life. Then, she wondered whether the driver of a trolley on course to run over five people could divert the trolley onto another track in which only one person would be killed.As it happens, when presented with these questions, most people agree it is not ethical for the surgeon to kill the patient and distribute her organs thus saving the other five, but it is indeed ethical for the driver to divert the trolley, thus killing one and saving the five. Foot was intrigued what the difference would be between both cases.She reasoned that, in the first case, the dilemma is between killing one and letting five die, whereas in the second case, the dilemma is between killing one and killing five. Foot argued that there is a big moral difference between killing and letting die. She considered negative duties (duties not to harm others) should have precedence over positive duties (duties to help others), and that is why letting five die is better than killing one.This was a standard argument for many years, until another philosopher, Judith Jarvis Thomson, took over the discussion and considered new variants of the trolley scenario. Thomson considered a trolley going down its path about to run over five people, and the possibility of diverting it towards another track where only one person would be run over. But, in this case, the decision to do so would not come from the driver, but rather, from a bystander who pulls a lever in order to divert the trolley.The bystander could simply do nothing, and let the five die. But, when presented with this scenario, most people believe that the bystander has the moral obligation to pull the lever. This is strange, as now, the dilemma is not between killing one and killing five, but instead, killing one and letting five die. Why can the bystander pull the lever, but the surgeon cannot kill the healthy person?Thomson believed that the answer was to be found in the doctrine of double effect, widely discussed by Thomas Aquinas and Catholic moral philosophers. Some actions may serve an ultimately good purpose, and yet, have harmful side effects. Those actions would be morally acceptable as long as the harmful side effects are merely foreseen, but not intended. The surgeon would save the five patients by distributing the healthy person’s organs, but in so doing, he would intend the harmful effect (the death of the donor). The bystander would also save the five persons by diverting the trolley, but killing the one person on the alternate track is not an intrinsic part of the plan, and in that sense, the bystander would merely foresee, but not intend, the death of that one person.Thomson considered another trolley scenario that seemed to support her point. Suppose the trolley is going down its path to run over five people, and it is about to go underneath a bridge. On that bridge, there is a fat man. If thrown onto the tracks, the fat man’s weight would stop the trolley, and thus save the five people. Again, this would be killing one person in order to save five. However, the fat man’s death would not only be foreseen but also intended. According to the doctrine of double effect, this action would be immoral. And indeed, when presented with this scenario, most people disapprove of throwing down the fat man.However, Thomson herself came up with yet another trolley scenario, in which an action is widely approved by people who consider it, yet it is at odds with the doctrine of double effect. Suppose this time that the trolley is on its path to run over five people, and there is a looping track in which the fat man is now standing. If the trolley is diverted onto that track, the fat man’s body will stop the trolley, and it will prevent the trolley from making it back to the track where the five people will be run over. Most people believe that a bystander should pull the lever to divert the trolley, and thus kill the fat man to save the five.Yet, by doing so, the fat man’s death is not merely foreseen, but intended. If the fat man were somehow able to escape from the tracks, he would not be able to save the other five. The fat man needs to die, and yet, most people do not seem to have a problem with that.Thomson wondered why people would object to the fat man being thrown from the bridge, but would not object to running over the fat man in the looping track, when in fact, in both scenarios the doctrine of double effect is violated. To this day, this question remains unanswered.Some philosophers have made the case that too much has been written about the Trolley Problem, and too little has been achieved with it. Some argue either that the examples are unrealistic to the point of being comical and irrelevant. Others argue that intuitions are not reliable and that moral decisions should be based on reasoned analysis, not just on feeling “right” or “wrong” when presented with scenarios.It is true that all these scenarios are highly unrealistic and that intuitions can be wrong. The morality of actions cannot just be decided by public votes. Yet, despite all its shortcomings, the Trolley Problem remains an exciting and useful approach. It is extremely unlikely someone will ever encounter a situation where a fat man could be thrown from a bridge in order to save five people. But the thought of that situation can elicit thinking about situations with structural similarities, such as whether or not civilians can be bombed in wars, or whether or not doctors should practice euthanasia. The Trolley Problem will not provide definite answers, but it will certainly help in thinking more clearly.
Different person may have different preference on the same feeling/sensation. In extreme case, some kind of pain might be preferred by some kind of persons, such as sadomasochists. Hence I concluded that there must be a deeper meaning than feeling which we should base our morality upon.
Here is another reading on trolley problem to check our ideas on universal morality.https://qz.com/1562585/the-seven-moral-rules-that-supposedly-unite-humanity/And here is how the trolley problem has evolved over time.https://www.prindlepost.org/2018/05/just-how-useful-is-the-trolley-problem/
Here is an example to emphasize that sometimes moral decision is based on efficiency. We will use some variations of trolley problem with following assumptions:- the case is evaluated retrospectively by a perfect artificial intelligence, hence no room for uncertainty of cause and effect regarding the actions or inactions. - a train is moving in high speed on the left track.- a lever can be used to switch the train to the right track.- if the train goes to the left track, every person on the left track will be killed. Likewise for the right track.- all the people involved are average persons who have positive contribution to the society. No preferences for any one person over the others.The table below shows possible combination of how many persons on the left and right tracks, ranging from 0 to 5.The left column in the table below shows how many persons are on the left track, while the top row shows how many persons are on the right track. \ 0 1 2 3 4 50 o o o o o o1 x o o o o o2 x ? o o o o3 x ? ? o o o4 x ? ? ? o o5 x ? ? ? ? oWhen there are 0 person on the left track, moral persons must leave the lever as it is, no matter how many persons on the right track. This is indicated by letter o in every cell next to number 0 on the left column.When there are 0 person on the right track, moral persons must switch the lever if there are at least 1 person on the left track. This is indicated by letter x in every cell below the number 0 on the top row, except when there is 0 person on the left track.When there are non-zero persons on each track and more persons on the right track than the left track, moral persons must leave the lever as it is to reduce casualty. This is indicated by letter o in every cell on the top right side of diagonal cells. When there are the same number of persons on the left and right tracks, moral persons should leave the lever to conserve resource (energy to switch the track) and avoid being accused of playing god. This is indicated by letter o in every diagonal cell. When there are non-zero persons on each track and more persons on the left track, the answer might vary (based on previous studies). If you choose to do nothing in these situations, effectively it shows how much you value your action of switching the lever, in the unit of difference of person number between the left and right track. This is indicated by question marks in every cell on the bottom left side of diagonal cells.
Pain is a distressing feeling often caused by intense or damaging stimuli. The International Association for the Study of Pain's widely used definition defines pain as "an unpleasant sensory and emotional experience associated with actual or potential tissue damage, or described in terms of such damage".[1] In medical diagnosis, pain is regarded as a symptom of an underlying condition.
https://www.prindlepost.org/2018/05/just-how-useful-is-the-trolley-problem/QuotePhilosophy can be perceived as a rather dry, boring subject. Perhaps for that very reason, divulgers have attempted to use stimulating and provocative thought experiments and hypothetical scenarios, in order to arouse students and get them to think about deep problems.Surely one of the most popular thought experiments is the so-called “Trolley Problem”, widely discussed across American colleges as a way to introduce ethics. It actually goes back to an obscure paper written by Philippa Foot in the 1960s. Foot wondered if a surgeon could ethically kill one healthy patient in order to give her organs to five sick patients, and thus save their life. Then, she wondered whether the driver of a trolley on course to run over five people could divert the trolley onto another track in which only one person would be killed.As it happens, when presented with these questions, most people agree it is not ethical for the surgeon to kill the patient and distribute her organs thus saving the other five, but it is indeed ethical for the driver to divert the trolley, thus killing one and saving the five. Foot was intrigued what the difference would be between both cases.She reasoned that, in the first case, the dilemma is between killing one and letting five die, whereas in the second case, the dilemma is between killing one and killing five. Foot argued that there is a big moral difference between killing and letting die. She considered negative duties (duties not to harm others) should have precedence over positive duties (duties to help others), and that is why letting five die is better than killing one.This was a standard argument for many years, until another philosopher, Judith Jarvis Thomson, took over the discussion and considered new variants of the trolley scenario. Thomson considered a trolley going down its path about to run over five people, and the possibility of diverting it towards another track where only one person would be run over. But, in this case, the decision to do so would not come from the driver, but rather, from a bystander who pulls a lever in order to divert the trolley.The bystander could simply do nothing, and let the five die. But, when presented with this scenario, most people believe that the bystander has the moral obligation to pull the lever. This is strange, as now, the dilemma is not between killing one and killing five, but instead, killing one and letting five die. Why can the bystander pull the lever, but the surgeon cannot kill the healthy person?
In the case of surgeon version of trolley problem, I think many people would make following assumptions that make them reluctant to make the sacrifice:- there is some non-zero chance that the surgery would fail.- the five patients' conditions are somehow the consequence of their own fault, such as not living a healthy life, thus make them deserve their failing organs.- on the other hand, the healthy person to be sacrificed is given credit for living a healthy life.- many people would likely see the situation in that healthy person's perspective.
Your objections seem to just be trying to avoid the issue. Let's assume the surgery carries no risks. The one dies, the others go on to live full lives. This is like assuming no friction in simple physics questions, or assuming the trolley will not overturn when it hits the switch at speed, explode and kill 20. Adding variables like this just detracts from the question being asked.
It is considered (rightly so) unethical to harvest a healthy condemned criminal in order to save the lives of all these innocents in need. Now why is that?
There is another solution: You have these 5 people each in need of a different organ from the one healthy person. So they draw lots and the loser gives his organs to the other 4. That's like one of the 5 trolley victims getting to be the hero by throwing the other 4 off the tracks before the trolley hits and kills him. Win win, and yet even this isn't done in practice. Why not? What is the actual moral code which typically drives practical policy?
Quote from: HalcIt is considered (rightly so) unethical to harvest a healthy condemned criminal in order to save the lives of all these innocents in need. Now why is that?I have some possible reason to think about.
- Perhaps the crime isn't considered severe enough for death penalty.
- Fear of revenge from the victim's relatives. There's always non-zero chance the secret will be revealed.
- Hope that there might be better options without sacrificing anyone, such as technological advancement.
- The lost of those five lives are not that big deal. Life can still go on as usual.
Millions of lives had died due to accident, natural disasters, epidemic, famine, etc. without anyone getting their hands dirty of homicide.
Quote from: HalcThere is another solution: You have these 5 people each in need of a different organ from the one healthy person. So they draw lots and the loser gives his organs to the other 4. That's like one of the 5 trolley victims getting to be the hero by throwing the other 4 off the tracks before the trolley hits and kills him. Win win, and yet even this isn't done in practice.In practice, that is a very rare circumstance.
There is another solution: You have these 5 people each in need of a different organ from the one healthy person. So they draw lots and the loser gives his organs to the other 4. That's like one of the 5 trolley victims getting to be the hero by throwing the other 4 off the tracks before the trolley hits and kills him. Win win, and yet even this isn't done in practice.
The cost/resource required could be high, especially if . Who will pay the operation?
The uncertainty of cost and benefit would make surgeons avert risks by simply doing nothing and noone would blame them.
Again, you seem to be searching for loopholes rather than focusing on the fundamental reasons why we choose to divert the trolley on a paper philosophy test but not in practice. I think there is a reason, but the best way to to see it is to consider the most favorable case, and wonder why it is still rejected. You seem to be looking for the less favorable cases, which is looking in the wrong direction.