"Spooner (the human detective interrogates the artificially intelligent robot): Human beings have dreams. Even dogs have dreams, but not you. You are just a machine, an imitation of life. Can a robot write a symphony? Can a robot turn a canvas into a beautiful masterpiece?
Sonny (the Artificially Intelligent Robot queries back): Can you?"
I, Robot (2004)
This dialogue from I, Robot is quite interesting. Spooner, the detective investigating a murder wants to distinguish humankind (or should I say human intelligence) from the artificially intelligent robot he is investigating. The argument he uses is something many would consider as the last line of defense against this rapidly developing technology: our creativity.
What if I were to tell you that the concept of creativity is not as old as you might think. Creativity, as it is understood today, is a recent concept that was popularized with The Art of Thought by Graham Wallis in 1926. Previously the word was used in reference to the divine creation. For most of human history, it was believed that only God could create something out of nothing. Humans merely replicated what God created. Modernity brought creativity to humans. Will the future bring creativity to AI?
For those that are skeptical of AI’s creative potential, it is already capable of writing music that we do enjoy listening to. Through experiment, we’ve proven that listeners, when uninformed AI produced the music, showed no distinguishable difference in preference between music written by AI or us. What is also interesting is that the same experiment also showed that this was not the case with lyrics. Lyrics written by AI simply did not speak to us in the same way.
For the moment AI is limited to recognizing patterns that have been pre-determined as creative. AI is already quite successful at replicating and reproducing these predetermined patterns into new patterns, a process that starts to become comparable to creativity. However, open-ended creativity and genuine design intent continue to be intrinsically human.
What AI has in its favour is that it can quickly draw needed information from vast databases. However, it is unlikely that a database will ever be able to cover each and every possibility. To give a simple example, the most up-to-date image recognition technology will have countless photo references of cats, yet, if a child draws a stick figure with some circles and lines, AI would struggle to recognize the image simply because it is not an image of a cat. AI is narrow. The system is only capable of working in limited number of scenarios and when it comes across an unknown dataset, it breaks down. Humans are able to make associations of things and ideas that could not have been associated before the original association was made. This is a fundamental aspect of the creative process.
As long as it is not possible to fully replicate the human mind with AI, there will always be space for humans in architecture. To further this thought exercise, let’s imagine it were possible to replicate the human mind, but even then there would still exist barriers between us. The human experience (therefore most of our knowledge) is entirely based on our experience with the world through our bodies. How would AI understand this if it does not have a body that it can experience the world with? A body limited by mortality.
Another interesting question is would we even be receptive of artificially intelligent creativity? Could there be another form of creativity? Sonny’s answer is very interesting as it highlights a potential reality that many would probably prefer to avoid. Over time, the bars we have set in order for AI to be considered intelligent have been continuously rising. Originally the Turing test, Jeopardy, then chess, Go, Natural Language processing, etc. The moving standard AI must meet before we consider it intelligent, is also a standard far too high for the majority of humans, if not all. The danger in this is that long before AI reaches this standard, it most likely will have already surpassed us. Do we also need to raise the bar for human intelligence and creativity?
Should we be afraid of the insidious rise of AI in architecture? There are many who are. The generalized argument against AI mostly revolves around the loss of jobs and the potential problems its rise may cause society. However, if we look at history, what we will find is that alongside technological development and the automation of labour, is also the continuous redefinition of our roles in the workforce. This redefinition of what we task ourselves with has kept us employed. You may argue that in the long run, this cycle may not be economically sustainable and will inevitably lead to the obsolescence of human labor, thus creating problems to the current economic model. (As a side note, perhaps what we should be doing is questioning our current economic model).
My personal view is that humanity has no good reason to want to keep doing the burdensome and mind numbing work of the past and should look forward to spending the precious time we have with tasks that are more fulfilling to both our inner and societal growth.
I think it is inevitable that AI will force us to question what it means to be an architect and designer. In this brave new world, we’ll need to reformulate the questions we are asking about the creative professions and let go of the fastidious tasks that do not require human creativity. When this happens, the traditional definitions of architecture and design will no longer be the same. I can imagine the architect / designer role transforming into a sort of emotional compass to a project. It is likely that in the foreseeable future our design decisions will continue to be made by, or at least verified by humans. We should continue to develop AI so that it can become an important tool that supports the creative process. When and if AI does become creative, surely more creativity in the world can only make it a better place.
A special thanks to Ismail Saadet and Rob Anderson whose interesting conversations helped inspire this blog.