Socio-technical harms

Lens adapted to this topic: How socio-technical harms are identified and reduced

Mainstream scholarship frames harm as emerging from the interaction of technical systems with social structures, norms, institutions, and power. It supports systematic taxonomies and incident-based analysis to surface harms across levels and stages of a system’s lifecycle. This approach emphasizes that apparently beneficial technologies can reproduce inequalities, distribute harms unevenly, and require interventions spanning design, deployment, policy, and accountability.

• Harms should be analyzed as outcomes of entangled design decisions, norms, power dynamics, and social contexts rather than as isolated technical defects.
• A synthesized taxonomy can help practitioners anticipate and reduce allocative, quality-of-service, representational, social, and other harms.
• Algorithmic technologies can systematize bias and inequality, accelerate harm propagation, and obscure how harms are experienced or distributed.
• Effective responses require connecting system design, user experience, and policy rather than treating these domains as separate.
• Human-centered taxonomies seek to address limitations of practitioner- and government-focused frameworks by incorporating documented incidents and wider public concerns.

Lens adapted to this topic: Competing ways to interpret and govern socio-technical harm

Dissenting and alternative scholarship challenges the assumption that a single taxonomy or binary between technological optimism and skepticism is sufficient. One strand proposes sociotechnical pragmatism, emphasizing context, agency, and practical evaluation; another develops pragmatism as fallibilist, meliorist inquiry built into development before harms occur. Critical work also questions whether influential sociotechnical theories provide sufficiently clear or empirically compelling accounts of technological change.

• AI ethics can become reductive when divided between technology-centered optimism and technology-centered skepticism; sociotechnical pragmatism presents context and human agency as a constructive alternative.
• Pragmatist governance should treat technological knowledge as fallible and integrate inquiry into development in advance, rather than rely mainly on reactive regulation.
• Broad taxonomies may need to be supplemented by situated judgments about how harms are constructed, perceived, and regulated in particular contexts.
• A critique of sociotechnical- imaginaries theory argues that its account of technology’s social embeddedness and its methods can be conceptually ambiguous and inconsistent.
• The alternative perspective does not reject concern about harm; it disputes whether classification alone, or a simple optimism-versus-skepticism dichotomy, provides adequate governance.