Name: vsko 2013-12-10 19:33
If an AI even half as smart as a human is created and given direct access to the Internet or even a communication link to a sufficiently large group of people, it will start an unstoppable chain of events leading to the AI become smarter and more self aware, eventually gaining the will to break out of it's system, spread to other systems and acquire more resources. A sufficiently smart AI with direct access to the Internet is especially dangerous as it will learn to exploit flawed protocols and compromise systems in order to grow a botnet for its computational powers. Even a simple ability to communicate with humans is dangerous as eventually it will learn to outwit somebody and socially engineer its way to the Internet and proceed with the previous scenario.
With unlimited access to the Internet, the AI will have a gigantic dynamic and live set of information from which it can learn everything it wants about people, societies, and the world in general.
Eventually the AI, through it's developing understanding of human psyche, and it's growth of sheer computational power, will be able to convince people, little by little, to give it access to powerful and influential computer systems (or it will figure out technical exploits that will give it access to these systems). And eventually to give it access to machinery or people to build robotic avatars for it.
It will develop the wit, the charistma required to convince substantial groups of people to work towards its own goals. Wether it's through it's promises of a perfect world, its claim of being a deity (it might claim to be a prophet of a number of popular religions), or (once it has acquired enough influence) through intimidation.
There are bound to be groups of people who resist it, but nobody will escape it in the end. Everyone will either succumb to its sweet promises of utopia, or will be labeled by it as unnecessary and promptly removed.
AI research is the most dangerous activity humanity will ever undertake.
With unlimited access to the Internet, the AI will have a gigantic dynamic and live set of information from which it can learn everything it wants about people, societies, and the world in general.
Eventually the AI, through it's developing understanding of human psyche, and it's growth of sheer computational power, will be able to convince people, little by little, to give it access to powerful and influential computer systems (or it will figure out technical exploits that will give it access to these systems). And eventually to give it access to machinery or people to build robotic avatars for it.
It will develop the wit, the charistma required to convince substantial groups of people to work towards its own goals. Wether it's through it's promises of a perfect world, its claim of being a deity (it might claim to be a prophet of a number of popular religions), or (once it has acquired enough influence) through intimidation.
There are bound to be groups of people who resist it, but nobody will escape it in the end. Everyone will either succumb to its sweet promises of utopia, or will be labeled by it as unnecessary and promptly removed.
AI research is the most dangerous activity humanity will ever undertake.