This paper argues that human rights impact assessments (HRIAs) in the technological sector must meet two fundamental challenges: assessing risks with no clear guidance in terms of the conduct that human rights law commands and an inadequate understanding of the effects technology has on society. The former is a legal problem; the latter is epistemological. While neither of these challenges, properly understood, are insurmountable, meeting them presents a few challenges. This paper explains why and offers possible strategies and a path forward, based on a critique of the business and human rights voluntary framework in which HRIAs are conceptually and genealogically embedded.