Neurodivergent Expertise in AI-Augmented Knowledge Work: Cognitive Debt, Reflective Scaffolding, and the Redesign Imperative
Citation. Ruttenberg, D. (2026). Neurodivergent Expertise in AI-Augmented Knowledge Work: Cognitive Debt, Reflective Scaffolding, and the Redesign Imperative. PsyArXiv. https://doi.org/10.31234/osf.io/yp4ru_v1
Summary. Ruttenberg offers a conceptual and methodological agenda for studying cognitive debt in AI-augmented knowledge work, with autistic and neurodivergent professionals treated as holders of design-relevant expertise rather than merely as a vulnerable subgroup. The paper defines cognitive debt as accumulated costs to sustained attention, learning transfer, and mental health when AI use chronically bypasses effortful cognitive engagement. Its main value for the wiki is not new causal evidence; it is a focused bridge between AI-mediated learning, cognitive-load/design questions, and individual-difference-sensitive work redesign.
Key claims
- Cognitive debt is framed as a time-accumulating function of attentional erosion, effort displacement, and affective depletion, indexed to individual and contextual thresholds rather than visible only through immediate performance decline.
- The paper treats learning transfer as the central performance risk: passive AI offloading may preserve short-term output while weakening the conflict-monitoring, problem-structuring, and effortful synthesis processes that support later unassisted transfer.
- Autistic and neurodivergent knowledge workers are predicted to show cognitive-debt indicators earlier and at lower AI-offloading intensities than neurotypical matched controls, not because of simple vulnerability but because narrower sensory tolerance windows, different precision-weighting profiles, and lived expertise in mismatched cognitive ecologies make unsustainable patterns more detectable.
- Reflective scaffolding is proposed as the alternative to passive offloading: AI should reduce extraneous burden while preserving core cognitive engagement through user-controlled pacing, sensory-aware interfaces, effort-preserving reading and synthesis tools, burnout-aware workflow management, and adaptive information flow.
- The paper proposes falsifiable longitudinal tests, including 16–24 week comparisons of passive-offloading versus effort-preserving AI use, threshold-transition detection in cognitive-debt accumulation, and earlier indicator timing among autistic and neurodivergent professionals.
Evidence & limitations
- The paper is a PsyArXiv preprint and conceptual/methodological advancement, not an empirical study of neurodivergent professionals using AI at work.
- The evidence base is explicitly assembled from adjacent literatures: general-population AI learning-transfer evidence, cognitive-offloading theory, autistic burnout, sensory processing, predictive processing, and participatory/design-justice work.
- The author acknowledges that there are currently no published empirical studies directly examining autistic and neurodivergent professionals’ interaction with AI tools in knowledge-work settings; the contribution is therefore agenda-setting and hypothesis-generating.
- The Drive PDF hash matched the public OSF download for the PsyArXiv DOI, supporting public provenance for the exact ingested file.