AI systems are becoming deeply embedded in every facet of our lives: from the way we listen to music, to the manner we are hired or assigned a credit card limit. Despite their wide acceptance, these systems are black boxes, of which the consequences, once they start interacting with the sociotechnical system, are difficult to anticipate and address. Design methodology is currently unprepared to deal with this issue. Existing methods are optimised for the design of non-transient tangible artefacts rather than for things that keep changing as they are used more. Therefore, we need new methodologies to address this challenge.
There is an explicit need to rethink the role of design in the AI development process. This project introduces a theoretical framework that supports designers in creating responsible AI systems by eliciting their potential unintended consequences (both positive and negative) and deliberately designing either for or against them.
We achieve this through the combination of abductive reasoning, (generative) prototyping, and framing. The framework is designed to be applied during the conceptual stages of the AI development process. As such, it places designers in a position where we not only work with the given material, but where we also have the opportunity to change it so that it will be more beneficial for society.
In order to showcase the potential impact of this theoretical framework on the conceptual design of an AI system, we partnered with a large multinational automotive company. Currently, one of the biggest obstacles for a large-scale electric vehicle (EV) adoption is the fact that most potential buyers do not have the needed infrastructure to charge their cars at home.
Therefore, they need to rely on public charging stations which can be extremely time-consuming, stressful, and uncertain. To lower the adoption threshold, we developed a conceptual design for an intelligent in-car system that works with its users to create charging routines that seamlessly fit within users’ everyday lives and the existing electrical infrastructure.