9

When a artist strokes their brush on a canvas and paints a beautiful work of art they may be referred as creative person. Or perhaps a musician or a writer. Does this creativeness come from the soul or from somewhere only a human could possess? Or will advanced robots be able to be programmed to pick up on the same patterns a human would and have the ability to create great works of art as well?

  • 4
    Related Wikipedia entry: en.wikipedia.org/wiki/… – user3164 May 10 '14 at 4:49
  • 2
    [Pick one] (a) Why not? Computers made of meat are creative. Why couldn't computers made of silicon be creative? (b) No, humans are special. We can do things computers can't. Meat rules! – user4894 May 11 '14 at 1:18
  • @user4894 -- false dilemma. I would assume that most who opt for what you call "(b)" do so not because they think meat rules but because they think humans are more than just the meat they are made of -- either through supervenience or some other means. – virmaior May 11 '14 at 5:46
  • You might find this interesting: people.idsia.ch/~juergen/creativity.html "Artists (and observers of art) get rewarded for making (and observing) novel patterns: data that is neither arbitrary (like incompressible random white noise) nor regular in an already known way, but regular in way that is new with respect to the observer's current knowledge, yet learnable (that is, after learning fewer computational resources are needed to encode the data)." – capybaralet Sep 12 '15 at 19:43
6

The answer is going to greatly depend on what you think creativity means.

If creativity is taken to mean "able to create something we find pleasant to experience", then the answer is clearly yes.

If creativity is taken to mean "able to engage in a creative process" that yields a novel creation, then much will hinge on what we think such a process is. Assuming that humans can engage in such a creative enterprise and this is what we talk about when we mean creativity, this becomes a strongly philosophy of mind question (which is not my specialization).

First, I want to comment on whether the human creative process can be identified with randomness. I am going to argue no for several reasons. First, little kids scrawl randomly on their drawing kits. And while we can say this involves some "creativity", we would say that the creative products of, say, John Howe are some how more creative. Or that a well thought Shakespeare play is more creative than a play generated by rolling the dice to figure out each line. Note that I am not saying that creativity could not include a component of randomness -- just that it is not randomness.

Second, creativity also seems to involve a type a freedom. And rarely is it good for this to be randomness. Randomness looks free, but it's doubtful this is what we want as freedom for creativity. I at least would be less rather than more impressed to learn an interesting painting was just the best pick of a litter of random scribbles than to think someone freely chose to produce it.

So then, take for instance, fractals. Fractals are pretty -- but they are not really creative at least not how I see the word nor are they generated according to a creative process that humans seem to use (they are generated through equations that lead to value fluctuations along the plotting axis and change colors as locations get more hits).

I would say that no computer so far engages in what I would call creativity, but I don't know enough to make a further answer in terms of if we understand creativity in the brain or the mind well-enough to guess as to how human creativty works.

  • 1
    If a computer were to understand methods of evaluating particular qualities (whether it be favorable circumstances in a game of chess or guidelines for what constitutes beauty) and the element of randomness presents various scenarios of how one might proceed from a given point - then a computer might be expected to select what it perceives to be an optimal random progression and act upon it. In this context randomness is not creativity itself - but it does serve to facilitate creativity. Of course whether the computer is blind to other great but sub-optimal options is down to code architecture – Avestron May 11 '14 at 17:37
  • "select an optimal random progression" is a phrase that defies meaning. If a progression is selected with some intent, then it ceases to be random and becomes a product of will. No computers yet exist that work in the way you describe vis-a-vis creativity. – virmaior May 11 '14 at 21:42
  • I have a feeling that I might have explained my meaning inadequately. Let us take chess as a simplification. Lets say that we present a computer with 3 random attempts to better its situation with a single king from square c3 and assume no hindrance. A king has 8 possible movements. If only three scenarios are 'looked' into and then evaluated then computer is able to choose optimally between the three randomly generated outcomes. This may be refined through the introduction of objectives ("move towards the center" = reduction in possibilities by 3-5). The comparison is digital vs analogue. – Avestron May 12 '14 at 5:19
  • @Avestron, I don't think you failed to explain yourself. Your chess analogy either shows you don't understand chess programs or you don't understand creativity. Computer chess programs are calculative -- not creative. They would never produce things like the early 20th century chess players did. They don't become more creative by having them roll the dice for 3 possibilities instead of 8. They don't take any chances. – virmaior May 12 '14 at 9:14
  • 1
    I would not pretend to claim and understanding of the underlying processes of the various chess programs out there. I shall also let slide the jab with regards my understanding of creativity - I consider the element of randomness to be an aide to creativity rather than to be creativity itself. I am certainly curious to hear which things you refer to in referencing the abilities of early 20th century chess players. Are we speaking of gambits, chess psychology or otherwise? Also people do face moments of indecision, perhaps forcing the choice through with little-to-no conscious bias. – Avestron May 12 '14 at 11:06
4

In a sense, computers are already capable of a degree of creativity. However it has always been humans who have been urging the computers to become creative - it has not been the computer that has engaged in the act of creativity for its own explorative or expressive requirements.

Creativity in computers has normally been based upon one or more elements of randomness - alternatively conformity to seed algorithms. In most cases creativity has been confined to the strictures of the rules coded into it by its programmers (and in the few other cases there are external forces at play - forces beyond the control of the computer (example - a gust of wind causing sprayed paint to divert towards one direction).

Existing forms of limited creativity include fractal graphics (graphical output on the basis of mathematical algorithms) - I am also certain that a computer's decision-making abilities allow it to perform a limited evaluation-driven pseudo creativity - such as when a computer projects suggestions to reduce drag over a surface area (based on the design inputted and the rules of aerodynamics for instance) or when a computer chooses its next chess move (particularly when its outside the library of opening moves (with their pre-defined values) and when it isn't caught up in a 'recognized endgame sequence scenario').


Even with the limitation of computers to abide within the strictures of the code set to them it is more than conceivable that as these strictures become ever more complex and nuanced, so too will the computer be able to expand its methods of 'thought'. These may become more and more life-like or we might be surprised to see that computers settle into a form of creativity that is unlike that of humans but undoubtedly creativity nonetheless.

Perhaps we will one day master the art of simulating in code the delicate processes of human learning, memory storage, interpretation, emotions, drives and ambitions, extrapolation, sympathy and desire, and more.

For a little (movie) inspiration I would suggest taking a look at (by no means a complete list), in no particular order:

Short Circuit (rather old), Bicentennial Man, The Matrix, iRobot, & Wall-E (animated)

  • 2
    I don't think randomness is akin to creativity... – virmaior May 11 '14 at 5:36
  • Randomness is an agent of chaos and creativity, besides being planned and derived, can be an accidental process also. One need not look too deeply into the human experience. One could argue that randomness is derived from the system clock or some other internal process - but as the actual process of randomization grows ever more distant from the seed factor(s) so too will it become less relevant. Especially if a computer grows to like certain digits or numbers over others (as humans tend to do). – Avestron May 11 '14 at 17:29
  • Liking certain things more than others is a process that appears random. It is not at all clear that it is random. – virmaior May 11 '14 at 21:40
  • Actually it is a process that is pretty clear. Individuals are (more or less) the sum of their derivations from their experiences. That one might favor the number 4 because it reminds him of the points of a compass, or because one has more good than bad experiences that one has associated with the number 4 is an exercise in discrimination against other possibilities in generating a random number from any set. Without randomness creativity is throttled and confined to the logical and familiar. – Avestron May 12 '14 at 5:30
  • 1
    @Avestron, virmaior -- I believe that you are talking past one another on the meaning of "randomness", esp. as applies (or doesn't) to mental processes. Many mental activities, creative ones especially so, are unpredictable (i.e. random, speaking colloquially). Maybe this unpredictability is only due to not being able to know all of the prior experiences that influence creativity; maybe there is intrinsic (physical/chemical) randomness in one's thought processes. A point that might be worth considering is whether this distinction makes a difference in the context of this question. – Dave May 12 '14 at 16:28
4

Computers have been shown to be able to generate classical music in a unique style.

Google Iamus, there was plenty of news coverage of it.

Iamus' Opus one, created on October 15, 2010 is the first fragment of professional contemporary classical music ever composed by a computer in its own style [2] (rather than attempting to emulate the style of existing composers as was previously done by David Cope). Iamus's first full composition, Hello World!, premiered exactly one year after the creation of Opus one, on October 15, 2011. Four of Iamus's works premiered on July 2, 2012, and were broadcast live[3] from the School of Computer Science at Universidad de Málaga[4] as part of the events included in the Alan Turing year. The compositions performed at this event were later recorded by the London Symphony Orchestra, creating the album Iamus, which New Scientist reported as the "first complete album to be composed solely by a computer and recorded by human musicians."[5]

Commenting on the authenticity of the music, Stephen Smoliar, critic of classical music at The San Francisco Examiner, commented "What is primary is the act of/ making the music itself engaged by the performers and how the listener responds to what those performers do... what is most interesting about the documents generated by Iamus is their capacity to challenge the creative talents of performing musicians"

Apologies for the long quote. One of the guys responsible for the research claims

its works are indistinguishable from human composers

I do not know what being creative could mean aside from creating art that passes as art, so I think the answer to your question is a resounding "yes"

1

I could name three arguments against this possibility.

  1. Objectives of any programming should be clearly determined, without ambiguities. Even if one uses a random number generator, its position in the code is not a random one and the programmer should know how, why and where use the generated numbers. Creation as objective has no such a determinacy. In many important cases - art, science or technology - this indeterminacy is due to the problem of the criterion - when is to stop and accept the result as some innovation? True innovation is a game-changer and it modifies rules that can be used in order to judge various values including its own. For example, when Planck had introduced a quantum hypothesis and Einstein had applied it in order to explain the photoelectric effect, this movement was without any value from the point of view of classical physics. However, what had happened from it resulted in a new physics, and now the initial innovation has a great value. Similarly, nobody could predict the appearance of jazz or rock music basing on the understanding of classical one, but they had changed the whole vision of the music. This means, that the rules do appear not early than the innovation appears. The programmer cannot in advance code those rules - they are absent.

  2. Who decides that the result is an innovation? I guess that the situation is looking like this: they switch on the computer, the code starts, then stops, writes down the result and they switch off it. Then they look whether the result is a creation or just a copy paste from the past. The "creator" does wish nothing, he has not initiated the creative process, he decides nothing and so on. In our practice, the creator starts the work, he decides when he can show it, he searches for recognition, and he promotes it and defends from critics. In the case of the computer creation, the programmers should perform all this authorship endeavor. This means that the creative work necessitates a conscious agency and a moral responsibility in order that the computer could insert him within the society and the latter accepts the author and its work.

  3. The computing is not a special kind of a physical process. In some sense, all physical processes, where one state follows from the previous one according to physical laws, can be considered as computing. So, why computers differ? I guess, there is an important anthropomorphic bias (or something similar) - machines' computing is very useful for us. Flowing river as a computing machine has no interest in our business, and we do not care about it and its complexity. We look in a special way on some physical entity because its functioning is useful. Why just by reason of this look this entity should transform itself in a conscious agent?

I think that the answer is "no, it's impossible". However, probably, all perception of the problem and its goals will change, so it could be changed to "yes". Women succeed in procreation of creative agents, why do not programmers?

  • Against argument 1 think that some computers can extract or discover rules depending not only on its code but on its "experience". Let's say that we have a military system designed to detect airplanes in photos. They do not try to code the rules but make a "learning" system where they "teach" it with learning photos: this is a plane, this is not... after a lot try an error the computer is able to detect planes that never were detected. It has "learned" the concept of plane. – borjab Oct 15 '14 at 10:30
1

In the paper Computing Machinery and Intelligence from 1950, Alan Turing argues that computers may be programmed to be creative.

  • 1
    What does it say about this site, when people's pseudo-philosophical musings get upvoted, and simple reference answers to what the giants had to say about a question get down-voted? – nir Sep 2 '14 at 2:43
0

It depends on two things. First one is the definition of creativity. And the second one is the question "depending on that definition, are the human creative?"

You can define creativity in two ways I guess.

  1. The ability to create something out of nothing.
  2. The ability to create something using out of something else and without any order to create.

If you define creativity in the first way, computers are not creative surely. But, are the humans? Does a painter creates his painting out of nothing or with the sense of art, with some education, with some emotions and mixes all those with a complex algorithm or does he just creates? That is a tough question, but I'll go with the first. I can't say that humans are creative in a way that they create something out of nothing but they just have the ability to analize complex things and then make some synthesis out of them ending up with something else. Thus, if you define creativity in the first way, both humans and computers are not creative.

So, if we follow the second definition of creativity, firstly I can say that humans are creative. Now, the question is, are the computers?

Before we started working on artificial intelligence, that was kind of an unrealistic dream. Just after we founded AI, it turned into a conspiracy theory, like "Computers are going to invade the earth, woaa!" However, now, as we know more about programming and its boundaries, this conspiracy theories are also gone.

But, that does not mean that computers are not creative. They are creative, but there is a limit of their creativity. I am going to follow some sports example instead of your painter example because it's more understandable.

Now, imagine a good football ( If you are American, I mean soccer, which is the real football ) defender. How can a defender be creative? What makes a defender good? Is it "he is never passed?" or "He is never passed the same way twice"? Just think about it. Even though you are a talented defender, if you not are experienced, avarage forwards would pass you easily. However, if you are experienced and avaragely talented, you can stop good players, because you can anticipate their moves better.

Now, that is what AI is. First, you define a defender. Then, you try to pass him with different moves and show him the result (good - bad). In the first time, he would just stop, because he doesn't even have any idea about what dribbling is. Then, he would start to move to stop you, but probably fail because he would not anticipate where you would go. But after a while, he would understand your moving pattern and start to stop you. That how good football games teach their players to play, both pre publish and after publish. With after publish, I mean, when you play agains AI player (CPU), when you try to pass a defender the same way 2 or 3 times, he does not buy it anymore.

Well, now, if we are done with our example, it is now time to make a statement:

Computers can be creative, yes. When you define a mission for an object in a program and write the code to help that object learn, it learns. The key word here is "mission". Objects learn according to their missions, in other word, their duties. A dutiless object does not learn, and an object does not learn something but its duty. So, is it different than human? I guess not. It's just the same with humans, we learn according to our duty. A painter's aim is to paint better, a defender's aim is to stop the forward, a carpenter's duty is to build a better chair. And just like a computer, they use both "already known" knowledge and experience.

So, computers are creative as much as human. And just like we are programmed by our DNA, they are programmed with code. Humans are stronger, because DNA is still more complex and amazing. But I still say, "still".

  • First, creativity as normal understood is not merely the power to create something (despite the linguistic semblance), so it's not really relevant whether the creation is from something or not. Second, the analogy with a sports game is completely flawed. There's nothing necessarily creative in being difficult to defeat or in being able to access all of the moves and use them well. Creativity invokes a certain type of insight absent from that. – virmaior Sep 2 '14 at 2:33
-1

You need to put things into perspective. Humans benefit from millions of years of evolution, which produces a being that doesn't start out creative or even intelligent at all, but has the ability to learn. And over the course of twenty years, surrounded by teachers, they learn, and some, but not all, become creative.

Currently, there is not very much effort put into the task to create thinking and creative computers, so you can't expect much today. With some serious effort, I'm quite sure computers could be made creative. There is on the other hand quite some danger involved with this.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.