AIGA 2022: how do we design AI systems that incorporate the end user

During a panel discussion on the critical intersection of AI and design practice at AIGA Design Conference 2022, Helen Armstrong, professor of graphic and experience design at North Carolina State University, describes the importance of empowering people to speak back to AI design systems.

It is a busy morning in Downtown Seattle as thousands of design enthusiasts descend on the city’s Convention Center for this year’s highly anticipated AIGA Design Conference. The American institution, which aims to promote and improve graphic design in the region, has put on an enormous selection of events for attendees to visit, from sessions on type design to understanding intercultural perspectives of the design process. Yet it was one of the smaller events – a panel discussion on understanding the links between AI and design – which really caught the imagination of the audience.

North Carolina State University’s Helen Armstrong gave impassioned plea to the hundred-or-so people there listening for designers to reconsider their relationship with AI and to allow consumers the opportunity to ‘talk back’ to it. “When I talk about AI,” Armstrong says, “I’m really talking about machine learning – we encounter it in our daily lives.”

But why should brand designers care? Well, in recent times we have seen a plethora of companies utilise AI technology to help consumers navigate the world (or their brands) in a better, simpler way. In 2019, for instance, McDonald’s bought software company (and AI specialists) Dynamic Yield in a move that would allow for the creation of intelligent menus which, store-by-store, would change based on data inputs. The compiling of data, so the idea goes, will mean McDonald’s can make better suggestions to customers in its drive-thrus, leading to an overall better experience.

“This technology has a lot of power,” Armstrong adds. This may not always be so great, she suggests, because a brand’s audience may disagree with what’s being suggested for them on a personal level. The point, therefore, is that designers must be able to incorporate the ‘rights’, as Armstrong phrases it, of consumers and allow them to choose for themselves.

An example of this would be Netflix offering a ‘thumbs up’ or ‘thumbs down’ button to its audience while watching a TV show or film. The brand offers us the chance to decide for ourselves what we like, as opposed to Netflix solely choosing on our behalf. These feedback loops may become crucial for brand designers to contemplate going forward as technology becomes even more integrated and complicated.

Armstrong argues designers can offer people these simple tools to ‘speak back’ to these AI systems. This may be more important than we currently comprehend because this technology has no consciousness – that is left to the designer instead. As Armstrong believes, the burden of conscience still falls squarely on the shoulder of the designer, not the technology. Should that be done, the possibilities from the merging of AI and brand design seem far more positive.