Opportunities and challenges of multidisciplinary algorithmic impact assessments

In recent years, technology assessments, evaluations, and audits have emerged as a key strategy for surfacing and managing technology systems’ societal effects. However, the evaluation of computational and algorithmic systems has largely been approached through a uni-modal and uni-disciplinary perspective that heavily privileges computer science and engineering disciplines. Silos of discipline and modality, we argue, impose limitations on conceptions of what constitutes ‘technology’ – the object of the assessment – as well as how to evidence its risks and impacts. Reflecting on our experience conducting a multidisciplinary and multimodal assessment, we consider the opportunities afforded by evaluating from various lenses at once, and discuss the steep challenges of navigating tensions between different ways of knowing. Ultimately, we call on assessors to welcome unconventional methods and underrepresented disciplines into the assessment space, as a means to broaden our political visions of mitigation and ethics integration.