Imagine a future where, for any task, AI has learned to create a maximally optimized interface – one that perfectly handles any scenario and delivers top results on the specified metrics.
No matter what you ask for, you get the best possible solution every time, calibrated mathematically and statistically.
But the price of this optimum is the loss of individuality. When everyone optimizes the same thing, the winning patterns get replicated and averaged out. Even if the interface adapts to each user, it will do so by the same rules.
It is precisely here that there may be a place for future interface designers – not only as reviewers of AI solutions.
A seemingly paradoxical task will now fall to them – to carry out a request not for improvement, but for a deliberate departure from the optimum for the sake of the product’s recognizability and character.
New visual languages were most often born as deviations from earlier rules. Of course, AI can create random deviations on its own. The human steps in as an author who controls the character and degree of such deviations: after all, unique does not mean inconvenient.
Still, at the cost of a few percent of efficiency, stepping away from the ideal makes a product better in ways that cannot be measured by speed and conversion: in emotions and the perception of lived experience.