Things Are Changing
There is no doubt things are changing and AI is developing by leaps and bounds.
However, design does not have to become an adversarial relationship between humans and machines.
If we actually retract from the traditional or as Maeda calls it, “classical” perception of design, where physical products and objects are created to satisfy a specific need, we’ll open ourselves up to the “commercial” design which builds innovation deep insights into UX. The pinnacle of this natural progression is “computational” design which uses data models, algorithms, and anything virtual to instantaneously satisfy billions of users.
Why AI Don’t Fly
Building an AI means emulating human behavior in different sets of contexts. The more variables, the more outcomes. At the core of this contraption lies the abundance of historical data on human behavior patterns.
The problem is there is no unbiased AI to populate itself with data
Every algorithm defining the conduct of AI is written by a person based on the data from other people. What if the biases towards race and gender make their way into the picture? This does not necessarily mean AI is inevitably flawed with someone’s bigotry but since machines are not blessed with their own judgment, they can’t make proper decisions.
With more learning and deep analytical capabilities, AI will deconstruct us to the point where some of the harshest historical biases concealed within the patterns of our language will play out in fresh colors.
Unfortunately, the impact a biased AI can have on society is greater than that. The problem of inclusiveness in design is a hard-fought battle for a lot of us and it only matters because we know how it feels. Feeling is not an AI prerogative and by delegating judgment or lack of thereof to a robot, we risk systematizing biases and discrimination into every physical or digital product we build with AI.