Sociotechnical harms of algorithmic systems: Scoping a taxonomy for harm reduction

MIT AI Risk Repository – Blog(Global) 16 Jan 2025 58

A structured harm taxonomy for algorithmic systems offers APS governance teams a reference vocabulary for risk assessment and policy design.

  • A 2023 peer-reviewed taxonomy classifies algorithmic harms into five categories and 20 subcategories across micro, meso, and macro levels.
  • MIT AI Risk Repository spotlights this as one of its indexed risk frameworks, making it more accessible to practitioners.
  • Published in 2023 and now featured in a repository blog - substantive but not a new or urgent development for APS readers.
  • Consider APS agencies developing AI risk registers or harm assessment frameworks could consider mapping the Shelby et al. taxonomy against existing departmental risk categories for gaps.
  • Monitor Governance teams may want to monitor which frameworks the MIT AI Risk Repository continues to surface, as the repository is increasingly cited in international AI governance work.

Implications are AI-generated. Starting points, not advice — see methodology for how they're framed.

View original source