Read more: 鈥�Existence special: Cosmic mysteries, human questions鈥�
OUR brains are incredible. They are the most complicated things in the universe that we know of. And yet there is no reason to think that they are anything other than flesh-and-blood machines 鈥� which means we should able be build machines that can emulate them.
Artificial intelligences on a human level would probably not remain at that level for long. AIs are expected to become smarter than us before 2050 (). A few researchers even think it could happen in the next decade.
Advertisement
For the first time, we would no longer be the most intelligent beings on the planet. The consequences could be stupendous. In 1993, the mathematician and sci-fi author dubbed this point 鈥渢he singularity鈥�, because he saw it as a turning point that would transform the world. So what will happen to us? Nobody really knows. 鈥淚t鈥檚 like cockroaches and dogs trying to predict the future of human technology,鈥� says Ben Goertzel, leader of OpenCog, an open source project to create AIs with general intelligence.
That hasn鈥檛 stopped people from considering various scenarios (). One distinct possibility is that AIs will exterminate us, which seems especially likely if the first are robots spawned in military labs.
Physicist and author David Deutsch of the University of Oxford has suggested that the way to avoid 鈥渁 rogue AI apocalypse鈥� is to welcome AIs into our existing institutions. But even if that were feasible, how could we compete with smarter and faster beings capable of working tirelessly 24/7 without ever getting tired or ill? They are likely to rapidly surpass all our scientific, technological and artistic achievements. Our precocious creations would soon end up owning the place.
One way or another, then, AIs look set to take over. One cause for optimism is that they will not be stuck on the planet like us fragile humans. A 1000-year trip to Epsilon Eridani is not so daunting if you can just turn yourself off until you get there. In fact, AIs may prefer to leave Earth. 鈥淭hey will probably work better in space, where it鈥檚 supercool,鈥� says Goertzel.
So we are not necessarily doomed to compete with AIs for energy and resources 鈥� a battle we are not likely win. With a galaxy to colonise, they may be content to let us keep our damp little planet. They might be as indifferent to us as we are to ants, or manage Earth as a kind of nature reserve.
鈥淎rtificial intelligences might be as indifferent to us as we are to ants鈥�
That might seem like a futile existence. But most people won鈥檛 be too bothered by the knowledge that they are inferior, Goertzel thinks 鈥� not as long as there鈥檚 sex, drugs and rock鈥檔鈥檙oll. Some people will continue to do science and art for the sheer joy of it, regardless of how poor their work is in comparison to the machines鈥�.
For Goertzel, the best case scenario would be that the AIs provide a 鈥渉uman reserve鈥� for those who want to stay as they are, while offering those who want it the chance to slowly transform themselves into something more than humans. 鈥淵ou would want it to be a gradual change, so at each step of the way you still feel yourself.鈥�
Stay human and die, or transform into a near-immortal superintelligence 鈥� what a choice.
Read previous article: 鈥�Existence: Will we die out?鈥�
Read next article: 鈥�Existence: What is the self?鈥�
