The rise of AI 'workslop' and organizational knowledge decay

Generative AI produces a large amount of junk. [1] According to a recent Harvard Business Review blog, employees are becoming increasingly reliant on this "workslop." [1] University of Oxford professor Matthias Holweg and analyst Thomas H. Davenport argue that when this occurs at scale, business processes and their outputs begin to deteriorate. [1] The experts refer to this organizational-level phenomenon as "knowledge decay." [1]

Holweg and Davenport identify verification, validation, and entropy as three key challenges organizations must address. [1] Verification involves disentangling authentic human content from AI-generated material that may contain glaring errors. [1] The time and critical thinking required for this verification can negate the initial productivity gains of using AI. [1]

In hiring, candidates are using AI to write resumes, tailor prompts to ranking algorithms, and secretly generate responses to interview questions in near-real time. [1] To counter this and avoid hiring subpar candidates, recruiters may need to spend more time conducting on-site interviews without AI access. [1]

The shift from productivity gains to "knowledge decay" suggests that unmanaged generative AI deployment creates hidden operational debt. The necessity for human verification—particularly in high-stakes areas like hiring, where candidates actively use AI to game ranking algorithms—indicates that organizations may need to revert to analog evaluation methods, such as on-site interviews, to secure authentic talent. The erosion of trust at the process level directly challenges the enterprise narrative that AI is a strict efficiency multiplier. What the sources don't address: How companies can systemically detect and filter this "workslop" before it becomes embedded in critical business workflows.