There’s No Such Thing As Artificial Intelligence

12 Jan

Because I couldn’t think how to start, I thought I’d open a book instead and look for inspiration. Who better to enlighten me about Artificial Intelligence than a movie director? “For motion pictures have a great deal in common with our own physiological and psychological processes – more so than any other medium.” Half-way there already! Just need to lift one more quote, scrape a little of the edge away, dip my brush in the film cement, and then splice them together.

“The opposite way is to find the one shot that serves as an introduction to a scene; the rest will follow naturally. Again there’s a grammar to it. Once you write your first declarative sentence, the narration flows.”  (John Huston, ‘An Open Book’)

Flows?

Flows…

A paint brush is not passive; it is the artist’s – or the decorator’s – accomplice. A good brush will take up the paint, offer itself to the surface, resolve form and texture.

That’s a good start – to something!

I remember now…. I somehow managed to miss out, from what should have been my definitive picture of embodied cognition (except as usual I got side-tracked), that the embodied school of thought maintains our intelligence is not confined to our brains but is much more evenly distributed throughout our bodies and what they come into contact with, so that we don’t need to do tremendously complicated calculations in order to cross the road but can rely largely upon the shape of our limbs and the movement of the scene in front of our eyes, finessed through continual small adjustments in tune with the flow of changing information, and, put together, that does a lot of the thinking for us.

Thinking for us…

If the artist’s hand has its own share of intelligence, why not the brush?

For this to be artificial intelligence, it must be both a product of our intelligence and adding something to it. Where do you draw the line between Natural and Artificial?

We just need to work out at what point nature becomes artefact and then today’s entry in my intermittent journal is done.

Is language natural or manmade, for instance? If we turn the clock back to Shakespeare’s time, to Chaucer’s Middle English, or the Anglo-Saxon of Alfred the Great, we can easily see how our language has evolved. The same is true of clothes, transport, paintbrushes, you name it. Clocks. All have evolved through use and at need. Is there anything that makes a word more natural than what it describes? Shoes and the words describing them are equally products of our minds, however snugly they fit nature. Can we not invent new words, if we so volish? Words are not physically inherited, like the shapes of faces, but learned in the crib and at our mother’s knee. If we happen to be born in a different country, we learn an entirely different set of words.

If I was a cheeky philosopher-type, I’d wonder what would happen if there was a machine that took messages passed under a door and returned them, after a short delay, translated into Chinese – but when the door opened, it turned out it was just someone who spoke Chinese!

Everyone we encounter who is able to process information that is meaningful to us, from our perspective is performing a function that, once put behind a curtain, can appear quite mechanical. Wherever our relationship with the people involved is impersonal, the communication indirect, the service factual, the difference merges into imperceptibility. Think about where the information really comes from when you ask somebody the time. The intelligence here resides less in us as individuals than in society in general; it can be manufactured all over the world using knowledge that was first stored in books written in languages that are no longer spoken. Much of the products of our own minds is ultimately of this kind, derived from elsewhere, and to every other mind the contents of our brain is largely only accessible via the artifice of language. If others to us provide this Artificial Intelligence, then so do we to them. If you follow this to its logical conclusion, it may be that is all we ultimately have.

Forget I said that. After all… There’s No Such Thing As Artificial Intelligence!

Words, we can certainly say, are not native to us, for all that they fill our heads.

Whenever we turn those thoughts of ours into words, we are indeed making use of artificial intelligence, as basic to us as picking up a stick and poking the world with it or, if its reach is not great enough, throwing it at the object of our attention. The intelligence spans us and the stick and is what we can achieve together.

Not that we’re the only animals capable of picking up sticks.

Does that mean we’re not the only animals capable of artificial intelligence, that it’s somehow a natural thing to do and that we’re just the best exponents of the art?

There are actually few skills we possess that don’t also occur, at however rudimentary a level, in some other animal (and countless skills they have that we can’t match).

But when it comes to Artificial Intelligence, we’re top dog and by a handsome margin.

No other animal has come close to building supercomputers, flying to the moon, or the self-service laundrette. At all these we are supreme – although admittedly not all of us have personally been to the moon, or even have access to a laundrette. There are parts of the world where clothes are washed by hand and supercomputers are still comparatively hard to come by.

Our ceaseless quest to make the world a little easier for ourselves will inevitably lead to some future supercomputer flying a washing machine to the moon and back without anyone having to lift a finger.

We’ll have put so much of our intelligence into the world that it will almost be able to make do without us. I suppose the only question left is whether this super-intelligence will still be regarded as ours.

The world is what we make it. We’ve always wanted it to be just a little more responsive to us: brushes that spring back to a point at the end of a painted curve, libraries with knowledge to spare, automatic braking systems on cars. We’re already surrounded by artificial companions, familiar machinery, and there will always be room for more. We won’t rest until we’ve succeeded in reinventing ourselves. We really are our own inventions.

If there is such a thing as artificial intelligence, it’s a big part of what makes us human. So far, that doesn’t seem to have ruled out inventing better and better ways to kill each other; on the other hand, that hasn’t stopped us successfully populating the globe, curing ills, building hospitals and desperately trying to figure out all the while how not to kill each other quite so often.

The great thing about progress is there’s always something new to look forward to, isn’t there?

While I was wondering in what direction to launch the second half of this essay – more pontification, more whimsy, a bit of both? – the Philae lander bounced down on a lump of rock 300 million miles away, the BBC describing its journey from the orbiting Rosetta spacecraft in terms of machinery not quite human and altogether more familiar:

“The robot probe, the size of a washing machine, was launched from the satellite on Wednesday and spent seven hours travelling to the comet.”

You couldn’t make it up, could you? Compared to my playful “prediction”, Philae both went miles beyond the moon and fell a long way short of complete autonomy. I expect the future will carry on being like that, too, a game of Snakes and Ladders.

Hello, Dave.

Hello, Dave.

And so the weeks went by…

Philae’s stuck in the shadows, somewhere, and today I’m still not sure what to say next.

You might ask: what’s it to be, then, utopia or dystopia?

Artificial intelligence saves the world, or allows the robots to take over?

Are these sensible questions?

Before we go any further – on this planet, or with these words – can we please stop talking of artificial intelligence as being something separate and start thinking of it as natural? If there’s a problem anywhere, it’s with us. No amount of clever problem-solving is going to avoid the age-old problem of human nature.

And if we can’t expect Artificial Intelligence to hold the answer to everything, neither in itself ought it become the problem – unless we value replacing ourselves above seeking better understanding.

What do we want for our world?

How important, how vital is it for us to take self-discovery so far as full mechanical replication of our selves? There are other ways of self-examination, less costly. Introspection only consumes a little of our time, increases the wear on our vanity.

If being more intelligent than a dog means you’ve become smart enough to start a world war, that does beg the question of what’s so admirable about our kind of intelligence.

Is more always better? And more of what?

As usual with these essays, I don’t know what I’m talking about, do I?

Intelligence.

IQ tests are supposed to tell you how intelligent you are, but you have to be fairly clever to understand them – and if you’re a dog you might as well eat the paper.

Let’s borrow that lame old excuse. Pretend the dog ate it and we have to start again without a definition of intelligence.

It’s something to do with living things – or, so I don’t contradict myself, to do with things like hands and paintbrushes that form part of a living system. Okay so far?

With only a sort of sketchy feel for what intelligence might be, we can fill in the gaps as we go, rediscovering it from how, in life, it gets used.

For a living system to sustain itself, it must be able to turn something else in its surroundings into a source of energy. *

This is where good intelligence is vital. Plants vie with each other to secure the best position in the sun. Animals go beyond that and move in on their food supply day or night.

A smart move is to regard other animals as your mobile larder. You wait for them to gather the energy for you, then pounce.

Humans are omnivores, happy to consume almost anything, even happier when they have a menu to choose from. That’s real intelligence. Not the printed word. The choice of what to eat next.

Since the first cave painters we’ve been trying to resolve our artistic urges with our hunting instincts.

Try as we might to redefine ourselves as higher, rational beings, strip us of our intellectual pretensions and you begin to see the grace behind the muscles at play, closing in for the kill. You can almost smell the blood.

It doesn’t do me many favours. I wouldn’t last a week if I had to fend for myself. A day. Hours. I don’t score too hot on IQ tests, either, but being beaten on an IQ test is less of a threat to me than being at the wrong end of someone (or something) else’s survival instinct.

I should be one of the losers, but for a further accident of birth.

Whilst intelligence extends the range over which we can control our environment, making sense of what’s going on within it until we can see where the next meal’s coming from, culture expands our control over space and time, becoming increasingly global.

Multinational corporations emblazon their success on the labels in our supermarkets, on satellite television and almost everywhere online. World food brought to a shelf near you with a seal of approval assuring us of our cultural superiority, whilst otherwise intelligent people in the least developed countries struggle to make a living, having already unwittingly before birth failed a test loaded against them by history.

The most enduringly successful example of Artificial Intelligence – if for a moment we may again acknowledge the concept – is probably money. Having enough of it in your pocket renders everything you say more interesting. Let’s see what’s in mine.

Cash

Smart Money

How interesting did that make me?

It started simply as an exchangeable unit of work, the use of which quickly became international. Early trade routes enabled work done in one country to earn food grown in another. However, the farther you get from the original place of work (and the people involved in doing it), money seems to acquire an anonymity, an abstract quality that offers almost limitless possibilities to those seduced by mathematics and the fact of possession.

Is there another way of looking at it?

Could it ever be a sign of intelligence to want to tilt the odds in someone else’s favour? If so, then Artificial Intelligence might not be such a meaningless idea after all.


(*) On Self-Organizing Systems and Their Environments (1960) by H Von Foerster http://www.alice.id.tue.nl/references/foerster-2003.pdf#page=14 – thanks to eavesdropping on the Twitter book club of @kvistgaard @jeffsussna @ruthmalan

More food for thought from @alan_winfield @Miles_Brundage @j2bryson and @ArtBourbon @FuturegenLabs and all, and cheers @mapbakery @b1ggd for the chat.

Advertisements

3 Responses to “There’s No Such Thing As Artificial Intelligence”

  1. Alan Winfield January 13, 2015 at 10:49 pm #

    I think it’s reasonable to ask what’s special about Artificial Intelligence, as defined by the Dartmouth conference in 1956, given that so many other human artefacts are both artificial and intelligence extenders. But your argument does seem a bit like those conversations that go: what is a car? Well it’s machine for personal transportation. Yes, but what is it *really*? Ok, it’s a clever construct of metal and rubber and plastics. Ah! so it’s just stuff like everything else. AI is, ultimately, just stuff like everything else. But I still think it’s useful to have a category of stuff labelled AI.

    • pabloredux January 14, 2015 at 4:34 pm #

      First of all, Alan, thank you for trying to discuss my wayward words.
      You’re right – of course there’s such a thing as Artificial Intelligence and I’m not really saying it isn’t a useful category. However, the pursuit of AI shouldn’t ever become an end in itself, it should always be seen in wider contexts. Your own work explores some of the boundaries of these.

      My aim generally in these pieces is to look at a topic from everyday perspectives that overlap with the more technical ones, without getting too technical (mainly because I can’t)! I may fail, but I am looking for a genuine overlap, although, ideally, I try to make it sound more far-fetched than it is – I confess!

      I admit I’m trying to produce something off centre rather than a neat argument, but I feel you’ve possibly gone astray in your “reductio ad absurdum” with the example of the car (to force me to explain myself)? Once we get to metal and rubber and plastics, haven’t we’ve lost sight of the idea of personal transportation?

      I don’t think I’ve argued, for instance, that intelligence can be assembled from paint brushes, but I have tried to suggest that over the years we have built just a tiny part of our intelligence into them.

      Here is the proposal from before the 1956 Dartmouth Conference, which defines artificial intelligence as being machine-simulated intelligence. However, it doesn’t really give a good definition of intelligence itself, instead listing varying ways the problem can be tackled. It doesn’t say why the project as a whole is a good idea or what its underlying purpose is, if successful:

      “For the present purpose the artificial intelligence problem is taken to be that of making a machine behave in ways that would be called intelligent if a human were so behaving.”

      This sounds remarkably similar to the ‘Turing Test’. It also leaves it open for every researcher to have had a different private purpose.
      Setting this aside, we still need to ask: what does constitute intelligent human behaviour?

      Rather than going for the dictionary, in the second part of my essay I tried to pick on an essential characteristic of intelligence. It’s unlikely I’m really going to do this in a few lines, but I wanted to use the attempt to play up an obvious aspect of our consumption of biomass – eating other animals – that we tend to play down when we talk about intelligence, to emphasise the effects of our intelligence on the rest of the world.

      You’ve recently signed “Research Priorities for Robust and Beneficial Artificial Intelligence: an Open Letter”, indeed your work is referred to in the accompanying document. I couldn’t help noticing that the reference below that of you and your colleagues was about “Causal Entropic Forces” by AD Wissner-Gross and CE Freer from 2013 – in summary:

      “Recent advances in fields ranging from cosmology to computer science have hinted at a possible deep connection between intelligence and entropy maximization, but no formal physical relationship between them has yet been established. Here, we explicitly propose a first step toward such a relationship in the form of a causal generalization of entropic forces that we find can cause two defining behaviors of the human ‘‘cognitive niche’’—tool use and social cooperation—to spontaneously emerge in simple physical systems. Our results suggest a potentially general thermodynamic model of adaptive behavior as a nonequilibrium process in open systems.”

      I don’t feel I was that wide of the mark with the one, rather older, reference I used in my essay (to defend myself against sensationalism), which discusses the role of entropy in self-organizing systems (including humans), observing: “By a self-organizing system I mean that part of a system that eats energy and order from its environment.”

      Overall, I tried to interweave a range of ideas into the whole piece that would give most readers at least one fresh angle – by no means all mentioned in this comment (I try to do my own thinking so it really does take me weeks to come up with this stuff, however dubious the end result is).

      However, the idea I wanted to leave the reader with was that when they next think about AI, it’s in terms of how we behave intelligently on the planet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Corona Radiata

Brains, Politics, Whatever ...

Emma C. Robinson

Researcher in computational methods for medical imaging; affiliated with the (developing) Human Connectome Projects; author of the Multimodal Surface Matching (MSM) software

Troovus

Social uniformity and sameness harass the individual most

Hack Circus Podcast

Behind the scenes of creativity

Chris Glynn

the thinking eye

Naturalistic Philosophy

Teleosemantically speaking

The Captured Thought

~The subjective experience of thinking

Utopia or Dystopia

where past meets future

i heart literati

trivia | brain science | a series of intrigues

Gravity and Levity

A blog about the big ideas in physics, plus a few other things

A HEAD OF DEPARTMENT’S BLOG

I'm maintaining this blog in my time as Head of English, University of Exeter. It's intended to reflect on matters relevant to colleagues and academics elsewhere: research, teaching, national policy, university life, managing our time, and so forth. All views are my own and not necessarily those of my employer, nor my department. If you like a post, please tweet to the world; if you don't please leave a comment to tell me why - Andrew McRae (http://humanities.exeter.ac.uk/english/staff/mcrae/; @McRaeAndrew).

Form Follows Function

All Things Architectural

The Dragon and The Phoenix

Bard's Journey into the Collective Subconscious and How It Shapes Our World

Praxtime by Nathan Taylor

The near future. The next 5-20 years.

George Zarkadakis

The official blog of author George Zarkadakis

Thomas Hurka

Professor of Philosophy, University of Toronto

Sean Enda Power

Professional work: articles, talks, public, interdisciplinary, and academic (includes work from 'Time and Illusion')

A Mind for Madness

Musings on art, philosophy, mathematics, and physics

Deep Safety

Victoria's musings on AI safety, machine learning, and rationality.

itakoyak

Living in Japan and posting some AI research thoughts.

Si Alhir (Sinan Si Alhir)

© 2016 Si Alhir (Sinan Si Alhir) — All Rights Reserved.

The Disorder Of Things

For the Relentless Criticism of All Existing Conditions Since 2010

Abeba Birhane

A little bit of Everything. Cognitive Science. Ethiopia. Ireland.

Writing my own user manual

Sergio Graziosi's Blog

Theory, Evolution, and Games Group

The EGG studies theoretical computer science, evolution, and game theory.

Filosofia: estudo e ensino

por Marcelo Fischborn

Charlie Alfred's Weblog

Thoughts on Software Architecture, Value Modeling, and Communicating about Complexity

Refugee Archives @ UEL

news and developments

The knowledge relation

Notes from a radical empiricist

Thom Scott-Phillips

Senior Research Scientist, Central European University, Budapest

Curious About Consiousness?

Me too! All thoughts belong to Heather Iriye, a PhD student in Psychology at the University of Sussex passionate about public engagement with science.

A Shared Madness

A FILM BLOG TO KEEP YOU SANE!

Nucleus Ambiguous

mind of science / science of mind

Samir Chopra

Refusing to Stick to the Subject

Sectioned

A blog about mental health & mental healthcare

Mind Hacks

Neuroscience and psychology news and views.

Bulldozer00's Blog

Exploring And Discovering

Weskaggs

Evolution: it's nothing personal

slehar

A topnotch WordPress.com site

Conscience and Consciousness

Academic Philosophy for a General Audience

Rolf Zwaan

Evolution: it's nothing personal

PhyloBotanist

Evolution: it's nothing personal

Wiring the Brain

Evolution: it's nothing personal

P-world and R-world

A solopsistic musing. Not terribly accessible.

%d bloggers like this: