Over a third of workers hoard expertise, fearing replacement by AI agents they train
agents training
| Source: Mastodon | Original article
Over a third of employees are withholding knowledge, fearing replacement by the AI agents they are tasked with training.
A new Fast Company survey finds that more than a third of employees are deliberately holding back expertise because they fear the very AI agents they are asked to train will eventually replace them. The study, published on 31 August 2026, reports that 35 % of respondents admit to hoarding knowledge, while 38 % say they are reluctant to teach colleagues in areas they consider personal strengths.
The findings echo concerns raised in earlier research on AI‑driven workplace anxiety. As we reported on 5 November 2025, the same pattern of knowledge hoarding was already emerging as companies accelerated AI adoption. The latest data suggest the behavior is becoming entrenched, with potential knock‑on effects for organisational learning and the performance of the AI systems they are meant to improve. When employees withhold tacit know‑how, AI agents receive incomplete training data, which can degrade their accuracy and limit the promised productivity gains. At the same time, reduced knowledge sharing hampers collaboration, slows innovation and may erode competitive advantage.
Stakeholders are now watching how firms will respond. Human‑resources and leadership teams are being urged to design incentives that reward open knowledge exchange rather than concealment, and to communicate clear pathways for reskilling that decouple employee value from the fear of automation. Analysts will also monitor whether the trend influences the rollout speed of AI tools, especially in sectors where expert input is critical for model fine‑tuning. Future surveys will be key to gauge whether interventions can reverse the hoarding impulse and restore the collaborative culture needed for AI to deliver on its promises.
Sources
Back to AIPULSEN