Tacit Erosion: Does AI-generated Feedback Undermine Critical Feedback Competence & Organisational Innovativeness? A Scoping Review
DOI:
https://doi.org/10.34190/ecmlg.22.1.5108Keywords:
Artificial intelligence, Feedback, Skill erosion, Human-AI interaction, Leadership, Organisational developmentAbstract
As generative AI systems become increasingly embedded in professional work environments, their implications for managerial and organisational feedback practices require closer examination. The technology designed to enhance individual and organisational productivity may be silently eroding a foundation of human development and organisational competitiveness. This paper introduces the concept of “tacit erosion” to describe the gradual and imperceptible weakening of feedback competence (nuclear dimension of critical thinking) that may occur under sustained exposure to AI-generated feedback. Feedback and specifically critical feedback have been linked to leadership development, organisational learning, and innovation capacity. AI-generated feedback is frequently characterised by availability, consistency, diplomatic phrasing, and encouragement-oriented responses. This may reshape expectations regarding the tone, immediacy, and emotional experience of feedback and influence how individuals tolerate criticism, interpret disagreement, and engage in direct interpersonal exchange. The erosion proposed here is labelled "tacit" for two reasons. First, the competence at risk is itself tacit: the ability to deliver, receive, and act upon feedback is practice-based, and relational, thus difficult to systemise. Second, the erosion occurs not through deliberate deskilling but as an unintended byproduct of routine interaction with AI systems. By imperceptibly shaping human critical exchange it makes the erosion of that competence difficult to detect. Following the PRISMA-ScR framework, this scoping review systematically maps literature from Scopus and Web of Science from 2022 to May 2026. It investigates emerging research on the effects of exposure to AI-generated responses and feedback in business environments and integrates four research streams: sycophantic tendencies in large language models (LLM), trust and feedback dynamics in AI-augmented teams, potential deskilling effects of AI use on innovation-related behaviour and risks to independent judgement in management learning contexts. Initial findings suggest contrast effects that make human feedback appear disproportionately harsh, reduced feedback resilience due to lower exposure to interpersonal criticism, and avoidance of direct feedback in managerial and team settings. This review contributes to knowledge management and leadership development research by systematically mapping the current publication landscape, identifying existing research gaps and by conceptualising the tacit erosion of critical feedback capabilities as a potential consequence of AI-augmented work. The review concludes with recommendations for practitioners navigating AI-augmented workplaces.