Archinect's Lexicon focuses on newly invented or adopted vocabulary within the architectural community. For this installment, we're featuring a term that featured heavily in our recent conversation with Felecia Davis on biases in artificial intelligence.

"Positionality" refers to the social and political context that creates your identity in terms of race, class, gender, sexuality, and ability status. Positionality also describes how your own identity influences, and potentially biases, your understanding of and outlook on the world. It is a critical concept in various fields, including cultural studies, gender studies, and social sciences, among others.

AI development is influenced by the positionality of the developers, and it can manifest in many ways. For instance, if an AI system is primarily developed by a particular demographic group, there's a risk that it might unintentionally favor that group due to implicit biases. Facial recognition systems, for example, have been shown to have higher error rates for people of color, potentially due to a lack of diversity in training data and in the teams developing these systems.

In AI development, positionality could also influence the choice of problems to be addressed and how solutions are designed and implemented. For instance, AI tools to predict criminal activity could be influenced by societal biases against certain demographic groups.

In building design, positionality can influence how structures are built and who they are designed for. Architects' own identities and experiences might shape their design perspectives, potentially resulting in buildings that are more usable and comfortable for certain groups but not others. A designer's understanding of accessibility, for example, may be limited if they do not have personal experience with disabilities or do not seek input from diverse groups.

Moreover, buildings themselves, being physical embodiments of cultural, societal, and individual preferences, also have a positionality. They can subtly or overtly signal who is welcome, reinforce societal norms, and even shape behaviors.

In both AI and building design, understanding and consciously addressing positionality can help in creating more inclusive, fair, and representative systems and structures. Efforts such as increasing diversity in teams, soliciting input from diverse user groups, and critically examining assumptions and biases are essential to these fields.

This article is part of the Archinect In-Depth: Artificial Intelligence series.