The impact of generative AI on knowledge work and organisational culture

Generative AI (GenAI) tools such as large language models (LLMs) are increasingly embedded in organisational knowledge work, raising questions not only about productivity but also about how they reshape everyday knowledge practices. This thesis examines how GenAI use relates to satisficing pressure, validation discipline, perceived retrieval quality (PRQ) of GenAI outputs, and knowledge sharing and reuse intentions. Building on theories of bounded rationality, human–automation interaction, information systems success and knowledge sharing, a process model is proposed in which GenAI use is expected to be positively associated with satisficing pressure (H1); higher satisficing pressure is expected to reduce validation discipline (H2); stronger validation discipline should increase PRQ (H3); and higher PRQ should foster intentions to share and reuse GenAI-assisted outputs (H4). Empirically, the study employs a pragmatist, deductive, cross-sectional survey design. Data were collected via an online questionnaire from 241 knowledge workers who had used GenAI for at least one real work task in the previous four weeks. The constructs were measured with multi-item scales, and GenAI use was operationalised via usage frequency and reliance (retention of GenAI-generated content in final work outputs). The findings partially support the proposed model. GenAI frequency is not significantly related to satisficing pressure, whereas reliance shows a small positive association. Satisficing pressure is negatively associated with validation discipline, but validation discipline does not significantly predict PRQ once reliance and perceived system reliability are controlled for. PRQ is positively correlated with sharing and reuse intentions at the bivariate level but does not exhibit a unique effect in multivariate models; instead, sharing and reuse intentions are primarily driven by GenAI reliance. Overall, the results suggest that GenAI’s impact on microlevel knowledge practices is shaped less by how frequently tools are invoked and more by how deeply their outputs are embedded in work products and how reliable they are perceived to be, with implications for the design of responsible GenAI use in organisational knowledge work.