Maths Predicts Computers Will Never Become Self-Aware, Saves Us From Judgment DayPublished by alistair on Thu, 15/05/2014 - 10:03
Unless the computers they used for their calculations are trying to throw us off the scent...
A scientific model of how consciousness works in the human brain might mean that it’s mathematically impossible for computers to ever become self aware. So, we’re never going to all get slaughtered/enslaved by robots! Well, not robots that think for themselves, anyway. I’m watching you, DARPA.
The study, “Is Consciousness Computable? Quantifying Integrated Information Using Algorithmic Information Theory,” was submitted to the online journal ArXiv on May 1. In it, Phil Maguire of the National University of Ireland in Maynooth and his colleagues build on earlier work suggesting that consciousness is based on the integration of information. When information becomes integrated, there’s a loss inherent in machine processes that Maguire’s team doesn’t think extends to actual consciousness.
Then, there’s a lot of complicated maths which predicts that computers will never be able to match the human level of information integration. Luckily, the maths works the same whether you can understand it or not, so you won’t need to go back to school to survive the robot uprising. And no, it’s not because you’re a beautiful and unique snowflake and your brain runs on hopes and dreams; the information integration necessary for consciousness is just so complicated that humans will never be able to decode how it works and bestow the ability on computers - according to maths, at least.
So, rest easy, everyone. It looks like our robot overlords may forever remain fictional… if this turns out to be true. Unfortunately, I’m not entirely sure it will. The model is predicated on the idea that the brain’s information integration has to be lossless. Speaking about the findings, Maguire said:
You have put in two bits, and you get one out. If the brain integrated information in this fashion, it would have to be continuously haemorrhaging information.
I can’t speak for anyone else, but I could make a rather long list of all the daft things I’ve done because of things I forgot (especially when drunk). I think “haemorrhaging information” would be a perfect way to describe how my brain works, so it’s probably possible to create a robot that has at least as much consciousness as I do. Hopefully that means it would also be as preoccupied with FIFA, Netflix, and Sheffield Wednesday as I am and wouldn’t bother anyone.
At least if it turns out this maths is wrong, you’ll probably be lasered to death by a smug robot before you even realise it.
(Thanks to the New Scientist for the info)