Thursday, May 28, 2015

From Smartphones to Cyborgs: Is Artificial Intelligence really ‘intelligent’?

Zombie apocalypse from The Walking Dead

(Image from IMDb)

In spite of movies like World War Z and TV series like The Walking Dead, there aren’t any organizations preparing for the zombie apocalypse.  

However, there are institutions preparing for the rise of the machines—that is, a time in the future when Artificial Intelligence (AI) becomes so advanced that it might develop a conscious “superintelligence” threatening the very existence of humanity.

Rise of the machines from Terminator 3

(Image from IMDb)

For example, The Future of Life Institute is a research organization preparing for the rise of AI and the existential risks it may entail.  In fact, this very year, 2015, the institute published an open letter warning about "potential pitfalls" that AI may hold for human survival.  

That letter has been signed by prestigious intellectuals, including physicist Stephen Hawking and AI researcher Nick Bostrom.  Lately, similar concerns have been expressed by business technology leaders such as Bill Gates.

(Image from IMDb)


(Image from IMDb)

The possible hazards of AI may no longer be mere science fiction, as the cinematic zeitgeist of our time seems to be telling us.  Plenty of upcoming movies such as Terminator Genysis and Ex Machina are warning about the perils of AI, which may, as The Economist has recently reported, move from science fiction to scientific fact sooner than we think.  Should we really be so worried?

Part of what make this question difficult is that it’s tricky to define what we mean by 'Artificial Intelligence.'  Heck, it’s challenging enough to define 'Intelligence.'  For example, there’s traditional IQ (Intelligence Quotient), but the power of the human mind also involves Emotional IntelligenceSocial Intelligence, and nowadays even Business Intelligence.

Perhaps it’s better, then, to speak of 'intelligences' in the plural, or varieties of intelligence.  Albert Einstein was certainly one kind of genius, embodying a scientific intelligence.  James Brown was a musical genius of artistic intelligence, and George Carlin a comic one of incredible witty intelligence.

So what kind of ‘intelligence’ is meant by the ‘artificial’ kind?  Perhaps we can clarify by differentiating human brains from computer hardware.  There are similarities for sure, but there are also major differences.  On one hand, both our brains and our computers share a common undertaking: processing information.  Our brains perform information-processing tasks; our computers work as information-processing technologies.  One the other hand, consider the differences:
  • Unlike computer hardware, brains are embodied: We humans come from the animal kingdom, and one major trait that separates us from plants and fungi (as any biologist will tell you) is that our evolutionary survival depends on bodily action (or sensorimotor activity).  Unlike trees and mushrooms, we grow muscles to move around and explore the world, and the brain is an organ that the animal body evolved over time to coordinate sense perception and muscular movement.  Many neuroscientists will point out that perception and movement are so closely linked in the brain that they are really one sensorimotor process.
  • Thus, mind is embodied: Since our brains evolved in the context of bodily action, our ‘higher’ mental functions (e.g., math and logic) naturally evolved out of our ‘lower’ sensorimotor activities.  For instance, our bodily movement naturally gives rise to what cognitive scientists call “image schemas,” or kinesthetic, visual patters, such as up-down, inner-outer, part-whole, etc.  We recruit these image schemas to create complex logic.  For example, in economics, we use ‘up-down’ schema to construct statistical logic—think of financial charts or graphs, where ‘up’ means ‘more value’ and ‘down’ means ‘less value.’

(There are other ways brains differ from computers—for instance, decision making in the brain depends on the body's emotional signals—what neuroscientists call somatic markerswhich allow humans to judge and evaluate facts.  I only highlighted a few differences here, but see Hubert Dreyfus’ classic book What Computers Still Can’t Do for more, or see my post on Embodied Cognition.)

Given these differences, is it right to call computers ‘intelligent’ machines?

At first, we might say no.  Intelligence, strictly speaking, is a quality of a conscious mind, and machines don’t have conscious minds per se.  But then again, this reaction is narrow.  For instance, strength is a quality of the muscles; buildings don’t have muscles, and yet we speak of 'strong' buildings.  The same goes for intelligence.  All computers are 'intelligent' in some sense.  That’s why it’s not a jump to refer to devices like smartphones.

So yes, we can say that AI is intelligent, at least in a limited sense of being able to execute certain tasks, even if it doesn’t have a conscious mind as we do.  (Computer scientists refer to this limited machine intelligence as “Weak AI,” which has no self-awareness.)  But will we ever have machines with conscious minds that can supersede our own?  (That is something computer scientists would call “Strong AI,” which just might have full self-awareness like human consciousness.)  Well, you'll need to decide for yourself, but I’ll briefly share my opinion.

Given what we know about embodied cognition, I’d say you probably have little to fear from your computer or smartphone (that is, Weak AI).  We know that a fully conscious mind depends on some form of embodiment—an autonomous, living body that's coupled to an environment (or what systems theorists call an autopoietic network).  But if that’s true, then when scientists combine Artificial Intelligence with Artificial Life to create something synthetic, such as a living cyborg or android, we may indeed need to prepare for the rise of conscious machines (or Strong AI).

In other words, a HAL Computer probably couldn't happen, but a Terminator Machine just might.


Don't worry about a HAL Computer 

from 2001: A Space Odyssey

(Image from IMDb)

But beware of the Terminator Machine 

from The Terminator

(Image from IMDb)

Or, for those who saw The Avengers: Age of Ultron, a robot like Ultron is unlikely, but an android like Vision may be a future possibility.


The robot Ultron

in Avengers: Age of Ultron

(Image from IMDb)

The android Vision

In Avengers: Age of Ultron

(Image from Wikipedia)

1 comment:

  1. There are times when it is definitely convenient.