No sabotage required. It’s typically poor tool with no strategy behind the so called deployment. AI snake oil salesmen claimed that AI could boost productivity AND be a scapegoat to cull the workforce, shareholders demand both, managers obliged, now the shit show is everywhere with no gains in sight.

Definition of sabotage in the article:

entering proprietary info into public AI chatbots, using unapproved AI tools, or intentionally using low-quality AI output in their work without fixing it

seems more of a design issue with the provided circumstances than active sabotage?