Artificial Intelligence is racing ahead with expectations that it will at some point be able to do anything that a human can do. Yet innately we do not think of computers as being creative – it’s almost the last bastion of our humanity and perhaps ego that there is at least something we are better than computers at. I want to address this question from several perspectives.
Firstly, we should define the nature of creativity. We can point to examples of creativity like painting, writing and scientific discovery. Creativity is something new and novel that also exhibits a characteristic of beauty – beauty as in art that has visual appeal or beauty as in a discovery that has intellectual appeal and/or usefulness. It is this feature of beauty that is hard to quantify – it typically requires a depth of understanding that goes beyond a simple analysis of the item and includes a deep context of culture and knowledge. Its thus likely that the true nature of creativity can’t emerge until we have AI that has a broader and deeper contextual understanding of an entire domain.
Nevertheless, most of what we do day-to-day is not particularly creative – we don’t often see the leaps of change that came about with major creative acts like Shakespeare and Einstein. We can hold the human race up to those high standards but the practical day-to-day creativity necessary to be human may not be as challenging.
Secondly, we only have two choices as to the nature of intelligence:
- Our brains are computational like a computer, in which case creativity is simply an emergent characteristic of complexity and scale.
- Our brains are quantum, in which case creativity may be dependent on the nature of quantum mechanics and fundamentally not computational. Thus, we may have to wait until quantum computing catches up to create a human equivalent artificial intelligence.
In either case it’s only a matter of time before we can create human intelligence artificially.
There are some interesting observations on the nature of intelligence. New ideas tend to emerge in our heads largely well-formed and complete – they don’t follow a structured methodology. We obviously need to think about a problem but many of the best ideas come while we are doing something completely different. Further there are techniques to create artificial challenges that tend to lead to more creative solutions (eg asking impossible questions). This hints at a model that is at least not sequential – either scenario-based testing or perhaps quantum in nature.
Thirdly we should look at the current state of the art in applying AI to creativity. There are lots of examples of art, music, science and maths being developed with the aid of AI. The important point though is that it is the combination of humans plus machines that drives the results.
AI can create a large array of interesting ideas but at this point it can’t assess which of these may possess the innate beauty that would move an idea into creativity.
Do you agree? Be interested to hear your thoughts on whether you believe AI could be creative, leave a comment below. Also if you haven’t already, explore our AI content hub for more insights.