Machine Intelligence = Machine Creativity?
Written for Interaction and Service Design Concepts, a course taught by Molly Wright Steenson at Carnegie Mellon University’s School of Design
Earlier this week, Google’s Artificial Intelligence company, DeepMind, announced that it built a new neural network that was able to navigate the London Underground. If you’ve ever visited London you’ll know this is no easy feat. But what made this story especially noteworthy is that it was the first time a neural network had been combined with memory. Meaning, what the machine learned by navigating the London Underground, it can now apply to other tasks, like navigating the New York Subway — without any prior programming data on the New York Subway System.
Unlike standard computer software that requires detailed instructions, rules and constraints, I believe that intelligent machines can acquire and apply knowledge in novel situations. Just like the human brain, intelligent machines look for patterns based on previous experiences. It can then apply those learnings to new situations and look for ways to become more efficient. For example, Google’s Artificial Intelligence (AI) will retain how it navigated the London Underground and can “reason” its way through other transportation systems (e.g. identifying train stops, schedules, connecting lines, etc.).
This capability to learn also highlights the key difference between traditional computers and intelligent machines. Traditional computers, which also access memory, can only solve problems or process data in a way that has been predetermined. So while a map application could tell you how to navigate the London Underground, it’s simply looking up the parameters that have been pre-programmed into the app and providing you with a result. But it wouldn’t be able to acquire new information (e.g. a map of New York City) and tell you the best way to get from Herald Square to Chinatown (FYI: Take the F Train to Grand Street).
This is just one example of how intelligent software is impacting our lives. It’s also being used to transform a variety of industries from insurance and banking to more creative tasks like screenplay development and industrial design. For example, industrial designers once relied solely on paper and pencil to diagram their inventions. Today they use digital programs like CAD. But we’re starting to see the next step of that evolution in intelligent software. In fact, researchers at Autodesk are actively exploring this concept through a new prototyping application called Dreamcatcher. With Dreamcatcher, a designer — with little to no engineering experience — inputs goals and constraints for the design and then allows the software to generate the optimal form. In the video below, we can see how designers and intelligent machines can work together as “colleagues, whose competence supplements your own will”[1].
While fascinating to watch, the video does pose some serious questions for me as a designer: How will this impact the role of the designer in the future? Can design be fully automated? Do humans have a place in the future of design?
While it’s easy to fear the unknown, I don’t believe that designers — or creatives in general — can ever be replaced entirely by intelligent machines. And that’s because I believe that design requires more than just intelligence. It requires skills that are innately human.
For example, I recently watched a short science-fiction film called Sunspring. What made this film unique is that it was written entirely by a machine. This computer was fed dozens of sci-fi screenplays and after analyzing them, it generated a new sci-fi script. While the computer technically “wrote” the screenplay, I wouldn’t consider it to be the screen play’s sole creator. The computer didn’t turn itself on one morning and feel inspired to write a sci-fi movie. A human being had that inspiration, motivation and the core idea to produce the screenplay. The origin of the work — the thoughts and emotions that inspire a human to create — are those innate human qualities that, I believe, a machine will never come to possess. As the producers admitted, “[the machine] exists somewhere between author and tool, writer and regurgitator”[2].
Furthermore, the movie also illustrated the challenges that machine generated content will face when it comes to the complexities of human communication. Just because a computer can string together words and sentences, it doesn’t mean that the result will make sense or that it can evoke emotions from the viewer. In the case of Sunspring, the script was hilarious — but only because it was completely nonsensical. Will the machines of the future be able to write a story line that’s (intentionally) funny or emotional? Can a machine understand what it takes to make a human feel something? Since even the most intelligent machine will always be operating based on set rules and patterns, they’ll fail to grasp what makes a good creative work or a breakthrough product (big ideas, surprise, novelty etc.).
Given the level of innovation occurring in AI, there’s no doubt in my mind that intelligent machines will play an increasingly larger role in the design process. However, I also believe that there will always be a place for a designer. Activities such as having the inspiration and motivation to put a new work out into the world, connecting with other human beings at a deep, emotional level and truly understanding their needs and desires — I believe these tasks will always be in the domain of a human designer. Therefore, these are the skills that we should look to cultivate, while always being mindful of ways in which we can leverage technology to handle tasks best suited for machines.
References:
[1] Man-Computer Symbiosis, J. C. R Licklider https://groups.csail.mit.edu/medg/people/psz/Licklider.html
[2] http://arstechnica.com/the-multiverse/2016/06/an-ai-wrote-this-movie-and-its-strangely-moving/
