I believe that for every human thought process, it is only a matter of time until it is understood well enough to be replicated by AI. That includes creativity. As Amir Panahi wrote, there are already many examples for research in computational creativity, like robots learning to draw in their own style. Personally, I'm keeping an eye on Interactive Storytelling and approaches to developing narratives with minimal guidance from a human author, so that the computer can keep the plot coherent and appealing while accomodating unforeseen user decisions.
You might argue that a computer system's "creativity" only consists of detecting patterns and reproducing them, maybe recombining them according to certain rules, maybe adding some random factor for exploration and some machine learning algorithm for improving based on the feedback of a human trainer...
But isn't this the same basic approach we humans use for being creative? Study famous artists, identify stylistic elements we like, recombine them to create our own style and ask other people for feedback so that we can improve? Deviate from the pattern we studied and the rules we have been taught by randomly changing some parameters and see if it works?
I think the current approaches using Genetic Algorithms, Reinforcement Learning and/or Deep Neural Networks are very promising for teaching AI to be creative in various tasks. I'm not sure about the time frame, but I'm confident that some day, AI will make decisions that the majority of humans considers "creative".
I believe using new techniques like deep learning, AI has progressed a lot in terms of creativity. Just check the IBM Watson cognitive platform that was used to create the first ever AI made movie trailer in 2016 for the movie called "Morgan", and that was just the start of it. There are lots of examples like AI-created music and paintings. Another example I can put my finger on is the recent news about Facebook AI programs that was shut down because the AIs were communicating using a "strange language created by the AI itself!" and they were not able to understand what they are discussing! These are only some of the examples. Deep dream project by google, Neural networks used in combined with AI, and lots of others projects are showing promising results. Considering what I've read and seen so far, I believe that AI is going to be able to make creative decisions in the near future.
This question is very general. Not so easy to say whether in all circumstances, AI is better than human. I do not agree that AI is unable to make the great decisions. Sometimes, AI performs better than human abilities, sometimes not. However, we cannot say that AI is unable to make the great decisions. I will give you some simple examples:
- Nowadays, almost nobody can overcome a chess-playing computer using AI which is changing its strategy all the time.
- Intelligent traffic light control is another system which decide to diminish waiting times before red traffic lights in a city in different situations.
By the way, AI is not limited to robots. AI is also using in data mining, machine learning, pattern recognition and so on.
I believe that for every human thought process, it is only a matter of time until it is understood well enough to be replicated by AI. That includes creativity. As Amir Panahi wrote, there are already many examples for research in computational creativity, like robots learning to draw in their own style. Personally, I'm keeping an eye on Interactive Storytelling and approaches to developing narratives with minimal guidance from a human author, so that the computer can keep the plot coherent and appealing while accomodating unforeseen user decisions.
You might argue that a computer system's "creativity" only consists of detecting patterns and reproducing them, maybe recombining them according to certain rules, maybe adding some random factor for exploration and some machine learning algorithm for improving based on the feedback of a human trainer...
But isn't this the same basic approach we humans use for being creative? Study famous artists, identify stylistic elements we like, recombine them to create our own style and ask other people for feedback so that we can improve? Deviate from the pattern we studied and the rules we have been taught by randomly changing some parameters and see if it works?
I think the current approaches using Genetic Algorithms, Reinforcement Learning and/or Deep Neural Networks are very promising for teaching AI to be creative in various tasks. I'm not sure about the time frame, but I'm confident that some day, AI will make decisions that the majority of humans considers "creative".
You might wish to read about Stephen Thaler's research with "Creativity Machines" to get a clearer picture of what is possible now, and what it may lead to.
AI can make decisions but whether it is right or wrong is based on the biased data we fed to the algorithm. MIT researchers explain about the potential dangers of biased data given to machine learning algorithm
Creativity to a whole new level, this question is so young and premature. Such questions are worthless to debate on when research is pacing at an alarming rate.
AI generally does not make decisions, because decision-making is a human's function. The way and quality of decision-making depends strongly on the EI. AI is an ordinary calculator, which, in addition to logical inference, can be configured to algorithmically extract knowledge from information. And recent investigations show, AI extract knowledge even better a human. Again - according to the algorithms that are determined by human.
About creativity. This characteristic of the level of consciousness, which is associated with intuition. Intellect is on the step below, even at the level of a human. Moreover, creativity is difficult to transfer into the language of algorithms.
to add more let me point out that ' The rapid deployment of intelligent automation is helping us set new standards of efficiency, speed, and functionality. Instead of being replaced, humans will see unprecedented job creation, as well as new opportunities to add more value'.