However, it leads me to thinking about a modest FOOM. What’s the least level of intelligence needed for a UFAI to be an existential risk? What’s the least needed for it to be extremely deadly, even if not an existential risk?
What’s the least level of intelligence needed for a UFAI to be an existential risk?
What makes you think that it takes a general intelligence? Automatic scientists, with well-defined goals, that can brute-force discoveries on hard problem in bio and nanotech could enable unfriendly humans to wreck havoc and control large groups of people. If we survive that, which I think is the top risk rather than GAI, then we might at some point be able to come up with an universal artificial intelligence.
Think about it this way. If humans figure out how to create some sort of advanced narrow AI that can solve certain problems in a superhuman way, why would they wait and not just assign it directly to solving those problems?
The problem is that you can’t make such narrow AI’s “friendly”, because they are tools and not agents. Tools used by unfriendly humans.
Luckily there is a way to impede the consequences of that development and various existential risks at once.
What we should be working on is a global sensor network by merging various technologies like artificial noses, lab on a chip technology, DNA Sequencing To Go etc.
Such a sensor network could be used to detect various threats like nuclear terrorism with dirty bombs, venomed water or biological pathogens early on and alert authorities or nearby people.
You could work with mobile phone companies to incorporate those sensors into their products. Companies like Apple would profit from having such sensors in their products by extending their capabilities. This would not only allow the mass production but would also spread the sensors randomly.
You might also work together with the government who is always keen to get more information. All it would then take is an app!
The analysis of the data could actually be done by the same gadgets that employ the sensors, a public computing grid.
This isn’t science fiction, it can actually be done. The technology is coming quickly.
And best of all, it doesn’t just protect us against various risks. Such sensors could be used to detect all kinds of health problems or stress levels and automatically call for help.
This isn’t science fiction, it can actually be done. The technology is coming quickly.
I was with you up until this sentence. Really, we can make a global sensor network today ? A network that would detect all conceivable threats everywhere ? This sounds just a tad unrealistic to me, though not logically impossible at some point in the future.
Thanks for working this up.
However, it leads me to thinking about a modest FOOM. What’s the least level of intelligence needed for a UFAI to be an existential risk? What’s the least needed for it to be extremely deadly, even if not an existential risk?
What makes you think that it takes a general intelligence? Automatic scientists, with well-defined goals, that can brute-force discoveries on hard problem in bio and nanotech could enable unfriendly humans to wreck havoc and control large groups of people. If we survive that, which I think is the top risk rather than GAI, then we might at some point be able to come up with an universal artificial intelligence.
Think about it this way. If humans figure out how to create some sort of advanced narrow AI that can solve certain problems in a superhuman way, why would they wait and not just assign it directly to solving those problems?
The problem is that you can’t make such narrow AI’s “friendly”, because they are tools and not agents. Tools used by unfriendly humans.
Luckily there is a way to impede the consequences of that development and various existential risks at once.
What we should be working on is a global sensor network by merging various technologies like artificial noses, lab on a chip technology, DNA Sequencing To Go etc.
Such a sensor network could be used to detect various threats like nuclear terrorism with dirty bombs, venomed water or biological pathogens early on and alert authorities or nearby people.
You could work with mobile phone companies to incorporate those sensors into their products. Companies like Apple would profit from having such sensors in their products by extending their capabilities. This would not only allow the mass production but would also spread the sensors randomly.
You might also work together with the government who is always keen to get more information. All it would then take is an app!
The analysis of the data could actually be done by the same gadgets that employ the sensors, a public computing grid.
This isn’t science fiction, it can actually be done. The technology is coming quickly.
And best of all, it doesn’t just protect us against various risks. Such sensors could be used to detect all kinds of health problems or stress levels and automatically call for help.
I was with you up until this sentence. Really, we can make a global sensor network today ? A network that would detect all conceivable threats everywhere ? This sounds just a tad unrealistic to me, though not logically impossible at some point in the future.