"Terminal Face": De-centering whiteness as approach to decolonize AI and Machine Learning

This paper presents the digital artwork/propositional design artefact "Terminal Face" (TF) and its diegetic affordances in (1) prototyping decolonized software (2) from within the white-western medium. We firmly reject the notion that any decolonizing can take place without a plurality of stakeholders, however what we are exploring in this work is an attempt of decentering-whitenessas-default in software. Given the absence of diversity in sites of code-instantiation we explore the production of actively anti-racist code through mitigating its bias(es). This work is an account of our effort to circumnavigate computational Eurocentristic-bias(es) through de-centering the privileged position of white ethnicities in AI/ML. TF’s proposed value is the exploration of user-identifiable skin-types and what we will call "meeting the code half-way" as alternative/supplementary pathways to code-justice. We document the context of emergence, technical considerations, reflections and implications of our design process and rationale.