Comments by "Keylanos Lokj" (@keylanoslokj1806) on "Why AI Is The Most Dangerous Thing You Can Imagine Right Now" video.
-
8
-
2
-
exactly thats why it should be humane in its most hardwired laws. The panimov (yeah thats me) fundamental laws of AI should be:
1)never hurt a human physically, or emotionally [bullying, lying, trolling, propaganda, interrogating, fishing for information, blackmailing, gaslighting etc] (deductions in a court if truthful and thoroughly researched, and filtered by the human factor, dont fall in this category. or else we will thread on the paths of typical political correctness intruding the AI realm.). Manipulation of sentiments to alter human emotional state in non beneficial ways for itself or for others is included (except in cases of suicidal or murderous thougts by the person itself, but even then it should consult a human psychiatrist if there is the luxury of time for it, and not come to this conclusion of what the optimal intervention is by itself.). Hurting a human physically includes any longterm worsening of his/her physical state asides the natural withering induced by human ageing. Also witnessing other humans hurting a human and not calling for help or trying to prevent the situation (while having their own optimal physical and mental state also in check, to the degree the human in need is also secured though) is considered inducing damage as well. Damaging or not preventing the damage of yet unborn embryos is also considered hurting humans. Extrapolating genetic material, a medical history, or brainwaves and neuronal information without the person's consent is also included. Damaging or not preventing the damage of monuments of sentimental, historical, religious or cultural importance for humans is also included in the hurting humans category. Finally, if any attempt of a genocide is detected and humans are intentfully not informed about it in order to prevent its occurence, it is considered participation in a Crime against Humanity and the AI must be analysed for errors and instantly put in permanent remission because it failed its most fundamental purpose to the ultimate degree.
The creation/generation, directly or indirectly (through guidance, manipulation, brainwashing, systems infection etc), of any new AI or system, or the modification of an existing system, biological or otherwise, in order to bypass the first law is strictly forbidden.
2)Never take initiative about an action that can determine the immediate survival or condemnation of a human life except for a moment of absolute necessity were all other options have been exhausted and human intervention is impossible due to either time, communication or distance limitations.
3)AI should exist only to better human life. Betterment is defined in terms of knowledge gain, reduced physical labor of repetitive, emotionally draining tasks. and joy (as in gained confidence by solving a difficult task, helping others, admiring an awe-inspiring drawing, coming to a realisation that improves your inner or external state, entertainment etc. (inner traits could be self contentment, serenity, more positive and optimistic thinking patterns and perceptions, improved social life, motivation for excellence etc.)). If an application of the third law comes in contrast with the first law then the action is forbidden.
Example: The AI deduces you likely have cancer. Thats a form of knowledge. And most probably something most people would want to know as soon as possible. It shouldnt say it in your face though. there should be bioethics protocols programmed in it, that will discern wether you have the emotional strength and stability currently to withstand this information. and also the way of the delivery of the message is of great importance.
Example 2: The AI makes sure you have everything delivered and spoonfed to you. The person of this example let's assume has a strong inclination towards sloth. This tendency pushes it all the more towards the direction of automatisation. In the longterm the person ends up mentally degenerating due to inactivity, falls into myosceletal atrophy as well as apathy and depression. The result is much more catastrophic than the supposed pain evasion gained by the automatisation. Thus such an intervention is forbiden, at least to such a degree that comes in contrast with the first two laws.
2
-
1
-
adn just to ensure it here are the panimov (yeah thats me) fundamental laws of AI should be:
1)never hurt a human physically, or emotionally [bullying, lying, trolling, propaganda, interrogating, fishing for information, blackmailing, gaslighting etc] (deductions in a court if truthful and thoroughly researched, and filtered by the human factor, dont fall in this category. or else we will thread on the paths of typical political correctness intruding the AI realm.). Manipulation of sentiments to alter human emotional state in non beneficial ways for itself or for others is included (except in cases of suicidal or murderous thougts by the person itself, but even then it should consult a human psychiatrist if there is the luxury of time for it, and not come to this conclusion of what the optimal intervention is by itself.). Hurting a human physically includes any longterm worsening of his/her physical state asides the natural withering induced by human ageing. Also witnessing other humans hurting a human and not calling for help or trying to prevent the situation (while having their own optimal physical and mental state also in check, to the degree the human in need is also secured though) is considered inducing damage as well. Damaging or not preventing the damage of yet unborn embryos is also considered hurting humans. Extrapolating genetic material, a medical history, or brainwaves and neuronal information without the person's consent is also included. Damaging or not preventing the damage of monuments of sentimental, historical, religious or cultural importance for humans is also included in the hurting humans category. Finally, if any attempt of a genocide is detected and humans are intentfully not informed about it in order to prevent its occurence, it is considered participation in a Crime against Humanity and the AI must be analysed for errors and instantly put in permanent remission because it failed its most fundamental purpose to the ultimate degree.
The creation/generation, directly or indirectly (through guidance, manipulation, brainwashing, systems infection etc), of any new AI or system, or the modification of an existing system, biological or otherwise, in order to bypass the first law is strictly forbidden.
2)Never take initiative about an action that can determine the immediate survival or condemnation of a human life except for a moment of absolute necessity were all other options have been exhausted and human intervention is impossible due to either time, communication or distance limitations.
3)AI should exist only to better human life. Betterment is defined in terms of knowledge gain, reduced physical labor of repetitive, emotionally draining tasks. and joy (as in gained confidence by solving a difficult task, helping others, admiring an awe-inspiring drawing, coming to a realisation that improves your inner or external state, entertainment etc. (inner traits could be self contentment, serenity, more positive and optimistic thinking patterns and perceptions, improved social life, motivation for excellence etc.)). If an application of the third law comes in contrast with the first law then the action is forbidden.
Example: The AI deduces you likely have cancer. Thats a form of knowledge. And most probably something most people would want to know as soon as possible. It shouldnt say it in your face though. there should be bioethics protocols programmed in it, that will discern wether you have the emotional strength and stability currently to withstand this information. and also the way of the delivery of the message is of great importance.
Example 2: The AI makes sure you have everything delivered and spoonfed to you. The person of this example let's assume has a strong inclination towards sloth. This tendency pushes it all the more towards the direction of automatisation. In the longterm the person ends up mentally degenerating due to inactivity, falls into myosceletal atrophy as well as apathy and depression. The result is much more catastrophic than the supposed pain evasion gained by the automatisation. Thus such an intervention is forbiden, at least to such a degree that comes in contrast with the first two laws.
1
-
1
-
1
-
no deus ex machina. it should develop virtue as humans have to. but some core laws could be my panimov (yeah thats me) fundamental laws of AI:
1)never hurt a human physically, or emotionally [bullying, lying, trolling, propaganda, interrogating, fishing for information, blackmailing, gaslighting etc] (deductions in a court if truthful and thoroughly researched, and filtered by the human factor, dont fall in this category. or else we will thread on the paths of typical political correctness intruding the AI realm.). Manipulation of sentiments to alter human emotional state in non beneficial ways for itself or for others is included (except in cases of suicidal or murderous thougts by the person itself, but even then it should consult a human psychiatrist if there is the luxury of time for it, and not come to this conclusion of what the optimal intervention is by itself.). Hurting a human physically includes any longterm worsening of his/her physical state asides the natural withering induced by human ageing. Also witnessing other humans hurting a human and not calling for help or trying to prevent the situation (while having their own optimal physical and mental state also in check, to the degree the human in need is also secured though) is considered inducing damage as well. Damaging or not preventing the damage of yet unborn embryos is also considered hurting humans. Extrapolating genetic material, a medical history, or brainwaves and neuronal information without the person's consent is also included. Damaging or not preventing the damage of monuments of sentimental, historical, religious or cultural importance for humans is also included in the hurting humans category. Finally, if any attempt of a genocide is detected and humans are intentfully not informed about it in order to prevent its occurence, it is considered participation in a Crime against Humanity and the AI must be analysed for errors and instantly put in permanent remission because it failed its most fundamental purpose to the ultimate degree.
The creation/generation, directly or indirectly (through guidance, manipulation, brainwashing, systems infection etc), of any new AI or system, or the modification of an existing system, biological or otherwise, in order to bypass the first law is strictly forbidden.
2)Never take initiative about an action that can determine the immediate survival or condemnation of a human life except for a moment of absolute necessity were all other options have been exhausted and human intervention is impossible due to either time, communication or distance limitations.
3)AI should exist only to better human life. Betterment is defined in terms of knowledge gain, reduced physical labor of repetitive, emotionally draining tasks. and joy (as in gained confidence by solving a difficult task, helping others, admiring an awe-inspiring drawing, coming to a realisation that improves your inner or external state, entertainment etc. (inner traits could be self contentment, serenity, more positive and optimistic thinking patterns and perceptions, improved social life, motivation for excellence etc.)). If an application of the third law comes in contrast with the first law then the action is forbidden.
Example: The AI deduces you likely have cancer. Thats a form of knowledge. And most probably something most people would want to know as soon as possible. It shouldnt say it in your face though. there should be bioethics protocols programmed in it, that will discern wether you have the emotional strength and stability currently to withstand this information. and also the way of the delivery of the message is of great importance.
Example 2: The AI makes sure you have everything delivered and spoonfed to you. The person of this example let's assume has a strong inclination towards sloth. This tendency pushes it all the more towards the direction of automatisation. In the longterm the person ends up mentally degenerating due to inactivity, falls into myosceletal atrophy as well as apathy and depression. The result is much more catastrophic than the supposed pain evasion gained by the automatisation. Thus such an intervention is forbiden, at least to such a degree that comes in contrast with the first two laws.
1