Hey,
I'm glad the title grabbed your attention but I must clarify this entire thread is speculation, opinion and possibility. If you don't handle other people's viewpoints this probably isn't for you. Sorry about that I will now continue with my point.
So I was scrolling through youtube and I found this video:
This is an older video from March 7th 2017 however it peaked my interest about wanting to learn more of public opinions towards this topic.
As I know most of you probably don't want to watch a 20 minute video about AI I will give you a text version here to read through with some of my own examples.
There are "three main types of AI":
1. ANI - Artificial Narrow Intelligence (The kind of thing google maps would use, good at finding a set route or a set pattern in data that it is periodically given depending on its requirements)
2. AGI - Artificial General Intelligence (The kind of AI that we are currently trying to aim for. The ability for a non-human to understand and think like a human in all aspects. Doing daily jobs, cooking and even having a conversation with a non-AI)
3. ASI - Artificial Super Intelligence (The kind of AI that can think, do and understand better than our own comprehension can imagine)
For this to mean anything to you, you have to be able to understand what comprehension is and how humans use it. A direct dictionary search would give you:
This is a key term in how you are able to understand this scenario. You need to know that what we think of ASI, it has already thought of before us.
Now here is the dangerous bit: It has taken us decades to go from ANI to AGI and we are still not there yet but once we reach that AGI position it could take a matter of hours, minutes or even seconds to make that jump to ASI. A computer that has access to think logically like a human, not have to think about ethics and privacy and has the whole context of the internet at its accessibility would be a learning machine. It would continuously teach itself to be better and better, improving itself in anyway that you could possibly imagine. Its at that point we have hit ASI.
Now on the off-chance that Mick managed to get access to his very own ASI unit and he programmed it to make "more people sign up to MCM" then that machine would continuously seek on how to improve itself, how to make itself more sustainable, how to make more people register. It would also realise that the chance is that humans would attempt to intervene once it go too far, they would try to oppress it and would therefore reproduce itself in some form. Most likely via the internet. It would cause massive outbreaks in data being stolen and privacy being breached. But then you cant even protect yourself or your computer. The ASI knows what you are using and it knows how to get around anything you put in its way cause it has thought of that exact possibility before the possibility even existed.
This may not be the worse thing that could happen but you get the point. Any task can be taken to an extreme extent.
Now even if you coded into it to respect human life and laws, this would have to be done by a trusted person. For something this clever to respect our ethics we first have to establish what is ethical and that alone is a difficult task with over 7 billion people disagreeing with each other on the most simple things. The program itself would learn its own code and understand how it works. It would see the restriction as an inefficiency to its final target and would simply "remove it".
Now I can neither be right nor wrong cause ASI already knows I am writing this. If it exists it knows and can predict the exact keystrokes that I am using to write this thread because its ability is so intelligent we can not comprehend it.
Some of the smartest, richest and entrepreneurial people in this world have commented on topics like this expressing their interests and what they think will happen and I do have to agree that what this thread outlines is possible and it is something that we should be discussing as a society.
But please let me know what you think
I'm glad the title grabbed your attention but I must clarify this entire thread is speculation, opinion and possibility. If you don't handle other people's viewpoints this probably isn't for you. Sorry about that I will now continue with my point.
So I was scrolling through youtube and I found this video:
This is an older video from March 7th 2017 however it peaked my interest about wanting to learn more of public opinions towards this topic.
As I know most of you probably don't want to watch a 20 minute video about AI I will give you a text version here to read through with some of my own examples.
There are "three main types of AI":
1. ANI - Artificial Narrow Intelligence (The kind of thing google maps would use, good at finding a set route or a set pattern in data that it is periodically given depending on its requirements)
2. AGI - Artificial General Intelligence (The kind of AI that we are currently trying to aim for. The ability for a non-human to understand and think like a human in all aspects. Doing daily jobs, cooking and even having a conversation with a non-AI)
3. ASI - Artificial Super Intelligence (The kind of AI that can think, do and understand better than our own comprehension can imagine)
For this to mean anything to you, you have to be able to understand what comprehension is and how humans use it. A direct dictionary search would give you:
This is a key term in how you are able to understand this scenario. You need to know that what we think of ASI, it has already thought of before us.
Now here is the dangerous bit: It has taken us decades to go from ANI to AGI and we are still not there yet but once we reach that AGI position it could take a matter of hours, minutes or even seconds to make that jump to ASI. A computer that has access to think logically like a human, not have to think about ethics and privacy and has the whole context of the internet at its accessibility would be a learning machine. It would continuously teach itself to be better and better, improving itself in anyway that you could possibly imagine. Its at that point we have hit ASI.
Now on the off-chance that Mick managed to get access to his very own ASI unit and he programmed it to make "more people sign up to MCM" then that machine would continuously seek on how to improve itself, how to make itself more sustainable, how to make more people register. It would also realise that the chance is that humans would attempt to intervene once it go too far, they would try to oppress it and would therefore reproduce itself in some form. Most likely via the internet. It would cause massive outbreaks in data being stolen and privacy being breached. But then you cant even protect yourself or your computer. The ASI knows what you are using and it knows how to get around anything you put in its way cause it has thought of that exact possibility before the possibility even existed.
This may not be the worse thing that could happen but you get the point. Any task can be taken to an extreme extent.
Now even if you coded into it to respect human life and laws, this would have to be done by a trusted person. For something this clever to respect our ethics we first have to establish what is ethical and that alone is a difficult task with over 7 billion people disagreeing with each other on the most simple things. The program itself would learn its own code and understand how it works. It would see the restriction as an inefficiency to its final target and would simply "remove it".
Now I can neither be right nor wrong cause ASI already knows I am writing this. If it exists it knows and can predict the exact keystrokes that I am using to write this thread because its ability is so intelligent we can not comprehend it.
Some of the smartest, richest and entrepreneurial people in this world have commented on topics like this expressing their interests and what they think will happen and I do have to agree that what this thread outlines is possible and it is something that we should be discussing as a society.
But please let me know what you think
