Synthocracy: The Worker Under Synthetic Evaluation
Hiring is only the first threshold. After the applicant becomes visible enough to enter the institution, a second transformation begins. The worker is no longer evaluated only by supervisors, colleagues, customers, craft standards, professional trust, or the visible quality of completed work. The worker is increasingly evaluated through systems that translate labor into signals: productivity metrics, task-completion rates, response times, message patterns, meeting behavior, customer ratings, delivery speed, sales conversion, error counts, compliance records, calendar density, platform activity, keystrokes, location traces, ticket resolution, warehouse scans, call sentiment, code commits, AI-generated summaries, and dashboard indicators. Employment becomes not only a relation between worker and organization. It becomes a relation between worker and rendering system.
The worker still has a manager. That is why the shift can be difficult to perceive. The manager may still speak, encourage, criticize, approve leave, assign work, conduct performance reviews, and make promotion decisions. The old hierarchy remains visible. Yet the manager’s view is increasingly mediated by dashboards. The worker does not arrive before the manager as a full person performing work in a complicated environment. The worker arrives as a surface of measurements: green, yellow, red; above target, below target; responsive, delayed; efficient, idle; engaged, disengaged; high performer, risk of attrition, low collaboration score, compliance concern, customer satisfaction issue. The manager may interpret these signals with judgment, but the signals prepare the field of interpretation before the conversation begins.
This is not entirely new. Work has always been measured. Factories counted output. Offices tracked hours. Sales teams counted revenue. Call centers measured handle time. Schools measured grades. Hospitals measured throughput. Logistics measured delivery. The human workplace has never been a pure zone of unmeasured craft. Some measurement is necessary. It can reveal overload, unfair distribution, safety risks, hidden effort, poor management, customer harm, and institutional inefficiency. A critique of synthetic evaluation must therefore avoid the lazy romance of unmeasured labor. The problem is not measurement itself. The problem is the elevation of measurement into the dominant reality of the worker.
A dashboard converts living work into measurable surfaces. It must do so because dashboards cannot carry the whole density of labor. They cannot fully contain the informal repair of a failing process, the emotional intelligence used to calm a client, the quiet mentoring of a new colleague, the judgment not to escalate a conflict, the decision to slow down because accuracy matters, the invisible work of preventing future problems, the social glue that keeps a team functional, or the moral hesitation that stops a harmful shortcut. These forms of work may be decisive, but they are difficult to render. What is easy to count becomes more visible than what is essential but resistant to counting.
Once rendered, the worker becomes manageable at scale. This is the attraction. The institution can compare teams, identify bottlenecks, reward output, detect anomalies, allocate resources, predict burnout, monitor compliance, optimize scheduling, and justify decisions. The dashboard promises fairness because it seems less personal than managerial intuition. It promises discipline because it makes patterns visible. It promises efficiency because it reduces ambiguity. It promises objectivity because numbers appear cleaner than human impressions. But numbers do not escape interpretation. They inherit the design of the system that selected them, the assumptions of the people who defined them, and the incentives of the organization that acts on them.
The worker learns this quickly. They learn which signals matter and begin to perform for the system that describes them. If response time is measured, they respond faster, perhaps with less thought. If ticket closure is measured, they close tickets more aggressively, perhaps while pushing unresolved complexity elsewhere. If call time is measured, they shorten conversations, perhaps at the expense of care. If calendar density signals engagement, they fill calendars. If messages create visibility, they message more. If customer ratings determine status, they manage feelings rather than solve deeper problems. If productivity software rewards visible activity, they remain visibly active even when real thinking requires silence. The worker adapts not only to the job, but to the model of the job.
This is one of the central economic distortions of synthocracy. When the description becomes consequential, people work toward the description. The metric ceases to be a window and becomes a target. The dashboard no longer merely reflects labor; it shapes labor. Workers begin to ask not only “What is good work?” but “What will the system read as good work?” The difference may be small at first, then decisive. A workplace that rewards machine-readable activity gradually produces workers who optimize for machine-readable activity. The organization may believe it is improving performance while quietly training everyone to serve the evaluation field.
The effect reaches far beyond office work. A delivery driver is evaluated by route adherence, speed, customer ratings, acceptance rates, GPS traces, and platform compliance. A warehouse worker becomes a stream of scans, movement, errors, and pick rates. A nurse may be seen through documentation completion, patient flow, medication timing, incident reports, and system compliance. A teacher may be rendered through attendance data, student scores, platform usage, behavioral notes, and automated reports. A customer service agent becomes sentiment, resolution time, escalation rate, and script adherence. A software developer becomes commits, tickets, pull requests, review time, and AI-assisted productivity measures. A salesperson becomes pipeline, conversion, outreach volume, forecast accuracy, and CRM hygiene. In each case, the worker’s lived work is not abolished. It is overlaid by a machine-readable double.
The double can be useful. It can protect workers from arbitrary managers. It can reveal hidden exploitation. It can document overwork. It can show that one person is carrying the burden others ignore. It can expose discrimination in assignments, pay, promotion, or customer treatment. It can help a worker argue for resources. But the same double can also become a cage. It may reduce the worker to what can be sensed, logged, and scored. It may punish forms of excellence that do not leave the correct traces. It may convert context into anomaly. It may make the worker permanently answerable to a version of themselves they did not author.
The emotional cost begins there. Under synthetic evaluation, the worker is not simply watched. The worker is interpreted. Surveillance produces anxiety because one is visible. Synthetic evaluation produces a deeper anxiety because one is visible as something. The worker worries not only that the employer can see activity, but that the system may misread the meaning of activity. A slow response may be carelessness, or it may be careful thought. A quiet day may be disengagement, or it may be deep work. A low customer rating may reflect poor service, or a customer’s anger at a policy the worker did not create. A reduced output rate may indicate laziness, or a hidden effort to correct defects before they multiply. The dashboard often sees the trace before it sees the reason.
Permanent visibility changes behavior even when no manager is present. The worker begins to inhabit the imagined gaze of the system. They wonder how inactivity appears. They worry about gaps, pauses, deviations, tone, message frequency, task status, route variations, and response delays. They pre-explain themselves. They document defensively. They generate evidence that they are working. They become careful not only in the moral sense, but in the performative sense: careful to leave the right trail. Work becomes partly the production of work’s own proof.
This proof burden is unevenly distributed. Workers in lower-trust positions are often rendered more intensely. Drivers, warehouse staff, gig workers, call center agents, retail workers, contractors, junior employees, remote workers, and precarious workers may face higher levels of monitoring because their standing is weaker. Senior professionals may be evaluated through outcomes and narratives; lower-status workers may be evaluated through traces. One class receives trust and interpretation. Another receives metrics and suspicion. Synthocratic evaluation therefore does not merely measure labor. It can reproduce class through visibility.
Informal context disappears most easily. Human workplaces, for all their unfairness, contain unofficial knowledge. A manager may know that a worker is caring for a sick parent, that a team has been understaffed, that a customer account is unusually difficult, that a project was saved by invisible coordination, that a delay prevented a worse error, that a quiet employee is stabilizing the group, that a decline in output is temporary and understandable. Dashboards can include notes, but they rarely carry the full moral texture of context. When the dashboard becomes the dominant reality, informal knowledge loses authority unless it can be converted into a field, tag, comment, exception, or adjustment. The worker’s life becomes admissible only where the system has made room for it.
AI summaries intensify this compression. A manager may receive an automated summary of performance, communication, meetings, tasks, customer interactions, or collaboration patterns. The summary may be helpful. It may save time and reveal trends. But it also becomes a second-order interface between manager and worker: not only data, but a machine-generated interpretation of data. The worker may be discussed through a narrative they never wrote, assembled from signals they did not choose, shaped by categories they cannot inspect. A summary can make the worker seem clear. That clarity may be the danger. The living person becomes administratively easy to understand.
The worker under synthetic evaluation therefore lives inside a field of self-optimization. They must do the work and manage the trace of the work. They must serve the customer and satisfy the metric. They must collaborate and appear collaborative. They must think deeply and remain visibly productive. They must be authentic and maintain the correct professional sentiment. They must refuse burnout while the system measures responsiveness. They must be creative while leaving evidence of progress. They must stay human inside an environment that prefers signals.
This does not require a cruel employer. Many organizations adopt these systems because they are overwhelmed, distributed, competitive, and hungry for coordination. Remote work, global teams, platform labor, compliance obligations, customer expectations, and cost pressure make measurement attractive. Managers themselves are evaluated by dashboards and may become intermediaries between worker and system rather than true governors of the workplace. The manager may also be rendered. Their team’s numbers define them. Their judgment narrows. Their patience becomes a cost. Their trust becomes harder to justify when the dashboard says otherwise. Synthetic evaluation moves upward as well as downward.
The result is a workplace where misreading becomes a permanent risk. A worker can recover from a human misunderstanding through conversation if the relationship allows it. But how does one repair a machine-mediated impression distributed across dashboards, reports, scores, alerts, and management reviews? How does one correct a pattern? How does one explain a metric without sounding defensive? How does one contest an AI-generated summary when the summary appears neutral and the worker appears interested? How does one challenge a productivity field that defines the terms of challenge? The worker may speak, but the dashboard has already spoken first.
This is why the worker is not only managed. The worker is rendered. Management gives instructions, evaluates conduct, and makes decisions. Rendering produces the version of the worker that management sees. It translates the living worker into operational surfaces before judgment occurs. Once rendering becomes dominant, the struggle over work moves upstream. The crucial question becomes not only whether the manager is fair, but whether the system’s picture of the worker is adequate to the reality of the work.
A humane economy under synthocratic conditions would have to defend workers not only from exploitation, but from reduction. It would require the right to know what is measured, how metrics are interpreted, when AI summaries are used, what consequences attach to scores, how context can be added, how errors can be challenged, and when human judgment must override measurement rather than merely decorate it. It would need to distinguish between measurement that supports work and measurement that captures the worker. It would need to ask whether a dashboard improves understanding or replaces it with a narrower reality.
The deepest injury of synthetic evaluation is not that the worker becomes visible. Many workers have suffered because their work was invisible. The injury is that visibility becomes one-sided, selective, and consequential. The system sees enough to judge but not enough to understand. It records enough to compare but not enough to witness. It produces a worker who can be managed more easily than they can be known.
The worker is not only managed. The worker is rendered.
