The Ubiquity of Artificial Intelligence
- Mark van Vuuren ©2017
Abstract
Computers have
come a long way in a short space of time. Software competes against human
endeavour, and has become stronger, e.g. chess and Go. Artificial Intelligence (AI) is now a
commonplace phrase, and produces creative output in, inter alia, literature,
poetry, music and art. It’s not unreasonable to assume that with time these
outputs will supersede human output in quality and quantity.
Introduction
In 2700 BCE Man was reliant on
tools for daily survival, such as the spear for hunting, tools for making and
containing a fire, and a tool for trade computations, i.e. the abacus.[i]
In the modern day, reliance on
tools has not reduced, and one wonders if existence without said tools is even
possible. When the handheld calculator was invented in 1967, little did anyone
guess future events.[ii] For example,
the square root of 4 is 2; the abacus confirms it. The square root of 2 is
1.414 and a bit. That ‘bit’ was calculated in 2010 to be 1 trillion digits, and
later calculated to 10 trillion digits.[iii]
It seemed unnecessary, but … Man is curious.
With regards the game of chess,
in 1996 World Champion Gary Kasparov lost to a computer, named Deep Blue.[iv]
This was a remarkable moment in computers, programming, AI (Artificial
Intelligence) and related words only to be found in OMNI magazine.[v] There
was talk that Man was well and truly on the back foot.
Since then chess software (known
as chess engines) has developed. The current World Chess Champion, Magnus
Carlsen, has a FIDE rating of 2835.[vi] The
top 3 chess engines have chess ratings as follows:
Computer Chess
Engines
|
Rating
|
Houdini 4
|
3277
|
Stockfish 6
|
3318
|
Komodo 9
|
3340
|
These chess engines can even play
each other, and have been doing so at the World Computer Chess Championship for
the last 22 years. The results of the last 3 years have been:[vii] [viii]
Year
|
Computer Chess
Engine winner
|
2014
|
Junior
|
2015
|
Jonny
|
2016
|
Komodo
|
And in further news, the 2500
year old Chinese board game Go, has a new champion. In early 2016, for the
first time, the engine AlphaGo beat the human then-champion Lee Se-dol.[ix]
Back foot indeed.
Communication
When driving in one’s vehicle and
using a GPS, a common feature is a voice instructing the driver when and where
to turn. These are pre-programmed verbal cues corresponding to an algorithm.
In a more complicated
development, there is a computer program whereby one types in any sentence, and
the software verbalizes what you wrote.[x]
This is known as text-to-voice. One can also instruct a page of writing to be delivered
through voice, even an entire text.
Getting more complicated, one can
verbally instruct one’s Smartphone, and receive audible confirmation, for
example, the intelligent assistant named SIRI, found in Apple products.[xi]
Crossing the Rubicon
In 1950 Alan Turing developed the
Turing Test.[xii] It is
a test of computer ability to present intelligent behaviour equivalent to, or
indistinguishable from, a human being. Put otherwise, if you are in a written
conversation with someone, can you tell if the other side is human or a
machine?
In 2014, a computer named Eugene
Goostman, passed this test, which no computer had passed before.[xiii]
There is currently a program
named CleverBot which allows one to type in questions and statements and
receive, what appear to be, coherent replies.[xiv] Try
it out.
There has been much development
in the realm of presenting a verbal statement and receiving a verbal reply,
like a conversation. These programs, in human mannequin form, are known as
androids, and even have names: Erica,[xv] Bina48,[xvi] Nadine,[xvii]
Jules,[xviii] Sophia.[xix]
This verbal riposte is
entertaining; one almost believes the banter is meaningful. At a certain point
of development it will be meaningful, probably first to one’s pets, then to
children, perhaps even to the uninitiated. Here’s the rub: It’s always other
people who are fooled, and the android will always be a clumsy automaton, or
will it? Man is curious.
A vital ingredient for AI complexity
is required: Creativity.
Literature
I writhed
with joy, which I experienced for the first time, and kept writing with
excitement. The day a computer wrote a novel. The computer, placing priority on
the pursuit of its own joy, stopped working for humans.
This is the final paragraph of a
short story submitted for the Japanese third Nikkei Hoshi Shinichi Literary
Award, titled “The Day A Computer Writes A Novel”.[xx]
What is notable about this competition is that judges are not informed which
novels were written by humans and which were “created” by human-computer teams.[xxi]
This short story, not a winner although it passed the first round of judging,
was robot-written.
To press a button and obtain a
manuscript exists even in the realm of non-fiction. Automatic essay generation
exists, whereby you input the title, specify topic keywords, word count,
research depth, bibliography, and within a few seconds the work is
auto-written.[xxii] [xxiii]
NaNoGenMo started as a casual
competition in 2013, with the requirement to write programming code during the
month of November that generated a 50,000 word novel. Now in its fourth year
it’s going from strength to strength. In 2015, 500 entries were received.[xxiv]
If software can identify aspects
of an author’s style, it’s possible, well, probable, that similar output can be
generated. This is the intent behind a Google project to predict and replicate
the works of famous dead authors, such as Dickens, Austen, Dostoevsky, Tolstoy,
Melville.[xxv] Sounds
kind of like a good thing.
Poetry
Given the following 2 poems,
which one was written by a human?
My adult oblivion[xxvi]
What muse and for
love thus removed it, upon
My hand, others
drag this thought to waft my
Name— drew. Oh,
and trembling terror
Brings him
fainting before thy path not have
Never fled, and
there to thought's, my boast, without
One on a ring of
men, three as finishing
Fate all bare
sense of thee how hastily.
Oh may keep one
thought of grave-damps o'er,
But only live the
palmer bent on my
Lovers gone, if
some palace-floor, diverse
From what
sweetness, which conceal'd chips called to
Me that floats
valley-fountain I wear too
Calm and absent
all, my adult oblivion
Yield, at once on
the kingdom of thy call
The second poem:
Leaning up[xxvii]
Is it and leaning
up the fruit lingers
In night with thy
love, I desperate and to
Take at my secret
influence thus to
The wise souls
stand further road will not say.
Let us to be
christ is changed was of thee,
And what a death
shall know thee withdraw their
Scheme lost
saints, and injury their ears. At
Our two
transform'd my foot’s dark, and precious,
And yet perplexed
into the lustre spent,
And all my
palm-tree something that censures
Life-angel
justified follows where
Nothing hindereth
o I may less truth
And such who in
armour mail'd green wave appears,
And sickness if
thou wert. Sonnet to do.
See the endnotes for the answer. To
make it simpler, which of the following 2 sonnets were written by a computer
program aka robot?
Sonnet 1[xxviii]
This is the same
harsh angle of the sun,
this is the same so deadly humid heat
I felt that week your ending had begun,
reflecting from the glass along the street,
the shattered bits of accidents or trash,
the careless cost of greed obsessed with speed,
the same damned world that made your system crash
and sprout that cancer like a roadside weed.
I walked those mornings to the hospital,
eyes downcast, sweating, breathing in the fumes
of fast Columbus traffic, senses dull,
or so I thought, but now this heat exhumes
the body of that grief. I saw, I heard,
and I remember, Mother. Every word
this is the same so deadly humid heat
I felt that week your ending had begun,
reflecting from the glass along the street,
the shattered bits of accidents or trash,
the careless cost of greed obsessed with speed,
the same damned world that made your system crash
and sprout that cancer like a roadside weed.
I walked those mornings to the hospital,
eyes downcast, sweating, breathing in the fumes
of fast Columbus traffic, senses dull,
or so I thought, but now this heat exhumes
the body of that grief. I saw, I heard,
and I remember, Mother. Every word
Sonnet 2[xxix]
The dirty rusty
wooden dresser drawer.
A couple million people wearing drawers,
Or looking through a lonely oven door,
Flowers covered under marble floors.
And lying sleeping on an open bed.
And I remember having started tripping,
Or any angel hanging overhead,
Without another cup of coffee dripping.
Surrounded by a pretty little sergeant,
Another morning at an early crawl.
And from the other side of my apartment,
An empty room behind the inner wall.
A thousand pictures on the kitchen floor,
Talked about a hundred years or more.
A couple million people wearing drawers,
Or looking through a lonely oven door,
Flowers covered under marble floors.
And lying sleeping on an open bed.
And I remember having started tripping,
Or any angel hanging overhead,
Without another cup of coffee dripping.
Surrounded by a pretty little sergeant,
Another morning at an early crawl.
And from the other side of my apartment,
An empty room behind the inner wall.
A thousand pictures on the kitchen floor,
Talked about a hundred years or more.
See the endnotes for the answer.
The issue is not how good the software is but rather how fallible we are.
If software can recognise, predict
and replicate the writing styles of famous deceased authors, one assumes this
may well extend to poets: Shakespeare, Yeats, Wordsworth, Wilde, even Homer. Sounds
kind of like a good thing.
Music
AI systems are able to create,
i.e. generate music, autonomously. Sony has released 2 songs, “Daddy’s car” and
“Mr Shadow”.[xxx] [xxxi]
[xxxii]
Global annual music sales are over $15 billion; this is a lucrative market.[xxxiii]
In 2016, Google’s Magenta program produced its first piece of generated
music. Just four notes were presented. After the creation the drums and
orchestration were added (by humans) for effect.[xxxiv]
Listen to it hear by clicking on the endnote.[xxxv]
If songs can be generated in an
instant and songwriters fees waived, and if software capability is a factor of
time and investment, then this area of development should reach a high level of
proficiency quite soon.
Perhaps, just as software can
recognise, predict and replicate writing styles of deceased authors, the same can
be said for music. Music in the style of Bach, Mozart, Beethoven, and even
popular music could be forthcoming. Sounds kind of like a good thing.
For example, there is a AI model
named DeepBach which can compose polyphonic chorales in the style of
J.S. Bach.[xxxvi] The
chorale is a rather formulaic piece of Lutheran church music that usually
reharmonizes a well-known melody.[xxxvii]
This is not a plagiarism of Bach, and it even fooled some critics; most
importantly, it produced genuinely
new work. Some critics have identified compositional errors like parallel
octaves, but the Bach-like patterns of characteristic cadences to the
expressive use of non-chord tones are accurate reproductions. The piece can be
heard at this footnote.[xxxviii]
It must be noted that music is a
powerful influencer, and not always in a good way, e.g. Seress’ Hungarian
Suicide Song, aka Gloomy Sunday.[xxxix]
[xl]
Furthermore, the perfection of a computer’s output is cause for concern: No
errors. Brian Eno warns of the same error-free, consistent production played
every time, with perfected voices having been run through pitch-correction
software; it’s simply not human.[xli]
The algorithm might need to be tweaked.
Art
What is art? (Perhaps an easier
question would be What is the meaning of life?) Art, by definition, is:
- The conscious use of skill and creative imagination especially in the production of aesthetic objects.[xlii]
- The class of objects subject to aesthetic criteria; works of art include paintings, sculptures, drawings.[xliii]
- The expression or application of human creative skill and imagination, typically in a visual form such as painting or sculpture, producing works to be appreciated primarily for their beauty or emotional power.[xliv]
The practitioners of art have a
different way of defining and describing art:
1.
"Art is either plagiarism or revolution."
Paul Gauguin, painter[xlv]
2.
"Art is a jealous mistress, and, if a man has a
genius for painting, poetry, music, architecture, or philosophy, he makes a bad
husband, and an ill provider." Ralph Waldo Emerson, essayist and poet[xlvi]
3.
"Art is meant to disturb. Science reassures."
Georges Braque, painter[xlvii]
4.
"Art is long, and time is fleeting." Henry Wadsworth Longfellow,
poet.[xlviii]
5.
"Art is the signature of civilizations." Jean
Sibelius, composer[xlix]
6.
"Art is a lie that makes us realize truth, at
least the truth that is given us to understand." Pablo Picasso[l]
7.
"Art is the Queen of all sciences communicating
knowledge to all the generations of the world." Leonardo da Vinci[li]
8.
"Art is essentially the affirmation, the blessing,
and the deification of existence." Friedrich Nietzsche, philosopher[lii]
These quotes are important; the artists
are accomplished, the statements reflect their experience, and it becomes clear
that art is essential to the human spirit.
Enter the realm of software, AI,
robots, etc. The term robotic art refers to any artwork that employs a form of
robotic or automated technology.[liii]
Example, brush strokes, complicated patterns applied with a unique instrument.[liv]
A combination of software and hardware that can produce an existing work of art
seems not dissimilar to a printer printing an image. Deep Dream Generator
receives images and re-interprets them to a style or medium, e.g. charcoal,
crayon, psychedelic, etc.[lv]
The key question, asked a bit fearfully, is Can a computer make art?
The Painting Fool is a
computer program that claims to answer this question by creating portraits based on its mood, assessing
its own work, and learning from its mistakes.[lvi] [lvii]
Its work has been exhibited in multiple galleries; there is even a Facebook
page.[lviii]
There is also the Robot Art
Contest which accepts submissions in two categories:[lix]
- Telerobotics, which is for robots that collaborate with humans (prize money of $10 000), and
- Fully Automated painting robots (prize money of $30 000)
French artist Patrick Tresset is
the inventor of Paul-IX, an automated sketch-bot that can outline a
still-life setting.[lx]
The robot comprises a camera, an arm holding a writing instrument, and software
that operates the process. The work has been exhibited in art galleries. There
is a recording on YouTube at this endnote.[lxi]
What’s notable about Tresset’s
robot is that the image is drawn on paper. Technology exists for an image to be
created … in air. Think hologram. Holovect is an application that draws
3D images in air using light.[lxii]
It’s a holographic display, presenting free-floating objects in space, and projections
can be manipulated in real time.
Hologram technology has developed
to the point that actual item A in Hologram Box A in Location A can present
image B in Hologram Box B at Location B; the Haptoclone does all, and
it’s in real time.[lxiii]
[lxiv]
Added to this is ultrasound energy which creates the sensation of touching and
being touched. In the public environment this might be used as a communication
device, seeing the person you’re speaking to in 3D, and being able to
experience the sensation of touch.
AARON is a computer
program that has been in continual development since 1973, and claims to create
original artistic images. Early programming was in C, then changed to Lisp. At
one point the programs ran on a DEC VAX 750 minicomputer. Given the early
development, many questions were asked about its output, if it was original,
the nature of creative work, the nature of creativity versus a program
following procedural instructions.[lxv]
Prisma is a program that
takes a digital image and changes it to look like a painting in the style of
particular painter. With this technology a simple image can have multiple
siblings, each with a different style (e.g. Picasso, Van Gogh, Munch, Levitan,
etc.). The software, in analyzing the pre-set artists even analyses their brush
strokes. This is a welcome approach in an art environment where emphasis on painting
techniques has waned, to be replaced by boldness or intricacy of ideas.[lxvi]
The interaction of art and
technology is even taught formally, e.g. the University of London
has a course titled Machine learning for musicians and artists.[lxvii]
Perhaps, just as software can recognise, predict and replicate writing styles
of deceased authors, this can be applied to art: Van Gogh, Vermeer, da Vinci,
Michelangelo, etc. Sounds kind of like a good thing.
Discussion & Conclusion
Moore’s Law came into effect after Intel
co-founder Gordon Moore noted that the number of transistors per square inch on
integrated circuits had doubled every year since their invention.[lxviii]
As a predictor, this Law was first scoffed at, and yet the growth rate of computing
power has been incredible.
What is required is a Moore’s Law for AI creativity.
The handheld calculator, the chess engine, passing the Turing test, the Go
engine, then literature, poetry, music and art. It’s difficult to establish the
rate of progress but it is evident.
One area of the AI field is the
philosophic questioning about a computer making art, what defines an
original piece of work, how human art appreciation differs from computer
art, etc. The philosophy of digital art is not an established field; if this field
interests you then refer to the work of Dominic Lopes.[lxix]
A connected area of AI art is the
programming, e.g. creating an original piece of work at the touch of a button,
or taking an image and reinterpreting it in a unique way. Some AI programs have
been mentioned, and although each seems somewhat limited, the way creativity
expands is to mix and grow exponentially on previous ideas.
A third area is the output of
art, which is the intention, after all. One might wonder when (not if)
computer art complexity will pass its Turing test. Let the public decide.
Two observations are notable:
- Program/ Application fluency is a factor of time.
- Creativity is pivotal for product development, and Man’s needs seem without end. Programmers and designers are creative in finding new applications. With software becoming a source of creativity, it will be very interesting to see what transpires.
The perseverant world turns,
motivated by profit and perceived value, limited only by Man’s imagination, until
now. Back foot indeed, and yet it sounds kind of like a good thing.
- end -
Further reading
- 25 real-life robots that will make you think the future is now[lxx]
- 5 Advanced humanoid robots you have to see to believe[lxxi]
- The 5 best computer chess engines[lxxii]
- These androids can hold a conversation and crack jokes[lxxiii]
- 9 computer-generated novels you should read, or attempt to, or at least look at in wonderment[lxxiv]
Keywords
AI, Android, Animatronics, Artificial
creativity, Artificial Intelligence, Chat robots, Chatterbots, Chatterboxes, Cognitive
computing, Computational creativity, computer vision, Creative computation, Creative
computing, Humanoid robot, Mechanical creativity, Robot, Robotics, Uncanny
Valley
Engines/ Robots/ Androids
referred to in this essay
- Deep Blue
- Houdini 4
- Stockfish 6
- Komodo 9
- Junior
- Jonny
- Komodo
- AlphaGo
- SIRI
- Eugene Goostman
- CleverBot
- Erica
- Bina48
- Nadine
- Jules
- Sophia
- Magenta
- DeepBach
- Deep Dream Generator
- The Painting Fool
- Paul-IX
- Holovect
- Haptoclone
- AARON
- Prisma
[iii]Shigeru Kondo calculated 1 trillion decimal places in
2010; https://en.wikipedia.org/wiki/Square_root_of_2.
Ron Watkins calculated the ten trillion digits; http://www.numberworld.org/digits/Sqrt(2)/
[xxvi]
Written by a robot http://botpoet.com/vote/my-adult-oblivion/
[xxvii]
Written by a robot http://botpoet.com/vote/leaning/