Let's be intelligent about the artificial intelligence uprising
This article was taken from the April 2015 issue of WIRED magazine.
Artificial Intelligence is not going to take over the planet and kill us all. Fear of the so-called AI singularity rests on the premise that computers will one day become self-aware because of Moore's law. This simplistic notion ignores just about everything we know about brains, neurobiology, consciousness and the nature of the mind. It is irrelevant that our computing technologies are indeed becoming more powerful every 18 months or so: they are, and forever will be, different from a self-conscious brain in form, structure and function.
When we talk about AI we mean computers capable of performing functions we attribute to humans: recognising faces, places and voices, learning from experience or communicating in natural language. We do not mean that the computer is aware of its functions thanks to an inner "I". Computers don't need to be self-aware to be more intelligent than humans, in the same way that a rocket does not need to be alive in order to outpace a bird.
So fear not AI, and do not confuse it with artificial consciousness. On the contrary, we should embrace AI as the technology that will provide the means to solve humanity's greatest challenges. The digital world began in the late 20th century, as small disparate islands of data processing and computation gradually coalesced, thanks to the internet, into ever-larger digital continents. This information evolution is now accelerating, as embedded computing transforms virtually everything into digital denizens of the internet of things: a deluge of data that offers humanity the greatest opportunity for a leap in progress since the invention of fire. But this data is useless unless we have an efficient means to extract new knowledge from it. The integrated information age needs technologies that will mine, correlate, contextualise and communicate new data insights.
The second crucial element is the computing systems' capability to learn. This is perhaps the most profoundly Promethean dimension of AI, because it will allow human/machine networks to learn faster from data than ever before. Today's deep-learning algorithms will evolve into the solution architects of tomorrow, working together with their human colleagues. This also means that our relationship with computers will change. From mere tools, they will become our partners and perhaps something more. Far from becoming mortal enemies, the intelligent computers of the future will become our trusted friends.