The rise of 'ghost work': Are employees unknowingly training AI systems that may replace them?

TEXT START: As AI systems become increasingly embedded into everyday work, employees across industries are beginning to question whether they are merely using artificial intelligence or unknowingly training the systems reshaping their own jobs.

The Dissection

This article is a legitimacy management document masquerading as investigative journalism. It accurately identifies the ghost work phenomenon—the invisible labor of employees training systems that will replace them—while systematically deflecting attention from the structural inevitability of what it describes. The article performs the journalism of concern without delivering the analysis of causation. It tells workers they should be worried, then offers them governance frameworks as a response to a mathematical problem.

The Core Fallacy

The article operates on a foundational error: it treats displacement as a governance failure rather than a structural feature.

The framing throughout is that if companies were more transparent, if workers had more consent, if governance frameworks were established, the tension between training AI and being replaced by it could be resolved. This is the exact institutional-fix delusion the DT framework dismantles.

The displacement is not a bug. It is the product. Ghost work is not an ethical problem companies must solve—it is the mechanism by which the post-WWII employment bargain terminates. The behavioral data, the workflow corrections, the institutional knowledge transfer: these are not regrettable side effects. They are the design.

The article's "What is changing, however, is the relationship between workers and technology itself" is the closest it comes to structural truth, and even here it retreats into "the central question is whether companies can build trust."

Hidden Assumptions

1. Worker agency is the variable that matters. The article assumes governance, transparency, and consent mechanisms can meaningfully alter the displacement trajectory. They cannot.

2. Organizational legitimacy is the battleground. The framing suggests the conflict is about trust, not survival. The real conflict is structural.

3. Augmentation is the alternative to replacement. The "shifting toward oversight, creativity, strategic judgement" language implies humans remain relevant at scale. The DT framework shows this is a temporary lag phase, not a stable equilibrium.

4. Displacement is still theoretical for most workers. The article presents it as "concerns emerging" and "growing sense." The CWA survey showing 36% of workers already believe they're training their own replacements is not a survey of future anxiety—it's a measurement of present reality.

5. "Most organisations still depend heavily on human judgement, creativity, ethical reasoning" is the final hedge, the soft exit, the permission slip to not face what the data already shows.

Social Function

This article is transition management theater—a sophisticated piece of copium that performs concern while containing the existential implications of what it describes. Its function is to:

• Acknowledge the ghost work phenomenon without triggering panic
• Offer governance frameworks as if they are solutions when they are actually containment mechanisms
• Preserve the fiction that workers and companies are negotiating the future when the math is already settled
• Keep the legitimizing narrative intact—that AI adoption can be managed ethically, that the employment relationship can be preserved through good faith

It is an article designed to be read by middle managers and HR departments and feel like they are "staying on top of the issue" while the structural transformation proceeds regardless of their awareness.

The Verdict

The article documents a terminal symptom—the growing worker recognition that they are training their own replacement—while treating it as a workforce trust problem rather than a structural consequence of the Discontinuity Thesis in action.

Ghost work is not the threat. Ghost work is the mechanism.

The threat is the permanent removal of productive participation from the employment model. The article's governance questions ("Should employees know when their work is being used to train AI? Should workers receive compensation?") are legitimate concerns within a framework that assumes the employment relationship remains viable. Under the DT framework, they are questions about how to manage the dying animal more humanely.

Workers who understand they are training replacement systems are not experiencing a misunderstanding that better transparency will resolve. They are seeing the future with unusual clarity.

The survival calculus is not about negotiating better terms for ghost work. It is about exiting the worker role before the math completes.