Design Futures in the Age of AI
(first published on LinkedIn Dec 13, 2017)
Eight years ago I wrote a design article titled Is Ux Becoming a Commodity? At that time I saw concerning trends that made me both nervous but also very excited. My excitement was driven by my desire to eliminate the exhausting production that was destroying my own creative process. It was not unusual for me and my peers to work long hours refining design blueprints, writing study protocols, designing screens as well as composing design reports. In short, I was tired; machines should be able to do this one day, I hoped. And for visual designers, the task was even more complicated. A lack of appreciation by less creative types was making it hard to enjoy the process of visual design. At the time, many of my older peers began to express high levels of discontent; some were leaving the field altogether and some moving to Product Management and or returning to school for new studies.
At the onset of the field of User Experience, it was rare to find a colleague who did not hold a graduate level degree, either a Masters and Ph.D., and there was an expressed level of discontent with the high influx of college-level designers with 2-3 year degrees quickly assuming comparable titles for half the pay. Roles such as Interaction designers, User Experience Design (UxD), Usability Specialist were being filled very quickly at a factory level intensity. Many practitioners in the field can attest to the messiness of design titles in Information technology space at the time. Also, companies seeking to hire “Designers” were also confused as to the many types of designers and their respective functions. I saw then an inevitable fracturing of Design, where there were also no standards of practice. Further to this fracture, I predicted that the output of designers (screen designs, information flow maps, and other deliverables) would be replaced by the production of some level of generative design models, building from learned human production over time.
Today application workflow creation is easily achieved by piecing together User Interface (UI) components with optional layouts and design appliques, which were once created by designers. More companies are looking towards UI component libraries to design applications or websites or optimise workflows. In fact, companies such as WordPress, Wix, Weebly, and Shopify would make manual webpage building a thing of the past. Libraries of UI components are readily available to be put together by development teams everywhere and at rapid rates.
The level of anxiety for Designers was well-founded. Over time, many of our functional deliverables such tasks as wireframes, workflows, interaction design flows, blueprints, web, as well as usability testing and focus groups begged for some level of automation. Personally, I secretly dreamt of ways to eliminate the manual churn so I could focus on more big design problems that required more mental engagement, as well as add value in ways that only human designers can add value. Almost ten years after writing this first article, I firmly believe we are once again at similar crossroads. We need to reset and begin to think about how to add real value in today's technological context.
Artificial Intelligence (AI) and Machine Learning (ML) are the new buzzwords; neural networks are learning more about how humans function including some elements of how designers design and also how users interact with computers. Machines are getting better at optimizing workflows, once devised by humans. In a recent chat with an ex-colleague who now works at Netflix, he mentioned that Netflix learning algorithms could auto-generate movie posters for higher levels of human acceptance, with very little input from UX designers. This form of generative design is not only limited to Netflix.
Earlier this year, I was only mildly shocked to read other similar headlines as mentioned by my friend at Netflix. Articles with headlines: AI And The Agency: Lingerie Brand Cosabella Replaced Its Agency With Artificial Intelligence, by Alison Weissbrot would be one of many such trickling. It seems inevitable that a dam is about to erupt. In more palatable examples, many social media apps can turn any user into an artist, with the many filters that can take regular pictures and make anyone feel like VanGogh. Less and less, Ux designers are called upon to do these tasks including the main staples of our functions such as conducting usability studies. Machines can now observe human behavior, in real time and compounded with many users’ patterns of usage can auto optimize just about anything. The fields of Machine Learning and Artificial Intelligence are the new fads unsettling the space of design.
Insights from Big Data is also replacing the further need for quantitative user testing, but we should heed caution and ensure there is designer human touchpoint and not relying solely on big data to inform design. In fact, there are many apps, such as Amplitude offering Insight tools with little need for the Psychological inputs of user researchers of the past.
I maintain the same ideas about Design erosion which I did in 2009 but perhaps with a more mature and confident view of that old position. In fact, I now believe the function of Design as we know it is at a critical point that requires a new way of thinking about the value we bring the Product development lifecycle. In the next few postings to come, I will focus on a few of these areas of value-add and propose some thoughts on how the function of design can evolve and provide even more tangible value to companies, especially in the changing world of information technology.