Algorithmic Management

In the AI Story

The genealogy is direct. Taylor's time-and-motion studies became Hawthorne Works efficiency experiments became Gilbreth's motion studies became the industrial-engineering methods codified in the twentieth century became the operations-research techniques of the fifties became the management-information systems of the seventies became the ERP deployments of the nineties became the algorithmic platforms of the 2010s. At each stage, the underlying logic — observe, measure, identify waste, redesign for efficiency, enforce compliance — migrated to new media while preserving its fundamental structure. The worker-as-system ontology Taylor articulated in 1911 became the operational foundation of platforms Taylor could not have imagined.

Amazon's picker provides the clearest contemporary illustration. The worker retrieves items from shelves brought by robots. Her movements are tracked by sensors. Her rate is displayed on a screen. Her breaks are timed by software. Her performance is evaluated against algorithmically determined targets, and she is disciplined for deviation from the prescribed pace. The system does not merely suggest the one best way — it enforces it through surveillance and consequence. The worker's knowledge of the whole (the supply chain she serves, the customers whose packages she assembles, the economic system her labor supports) is irrelevant to her role. She performs a fragment. The system performs the whole. The separation Taylor proposed as principle has been realized in its purest form.

The extension to knowledge work is what makes algorithmic management directly relevant to the AI transition. Tools that monitor keystrokes, measure time spent in applications, score collaboration patterns, and evaluate output through automated metrics bring Taylor's framework to domains his century could not reach. The knowledge worker augmented by AI can be tracked, scored, and managed through algorithmic systems with a precision Taylor would have envied. The Berkeley study documents the specific pathology: workers who adopted AI tools worked faster, took on more tasks, expanded into new domains, and found the freed time immediately colonized by additional work. The tool made more work possible; the organizational culture, still operating on Taylorist assumptions about the relationship between output and value, converted that possibility into expectation.

The alternative is visible in Segal's account of the Trivandrum training, where AI deployment was accompanied by deliberate cultivation of judgment rather than algorithmic enforcement of output. The tool is the same. The infrastructure around the tool differs. The organization that deploys AI through algorithmic management produces intensified components. The organization that deploys AI through amplification produces conductors. Both responses are available. Only one is consistent with the inversion the tool makes possible.

Origin

The term gained currency in labor-studies scholarship beginning around 2014, originally to describe the management of Uber drivers and platform gig workers. Academic consolidation came through the systematic reviews of the early 2020s, notably the 2023 Management Review Quarterly synthesis by Jarrahi and colleagues. The phenomenon's intellectual lineage runs from Taylor through the industrial-engineering tradition to the contemporary platform economy.

Key Ideas

Computational scale. Algorithmic management applies Taylor's methods at speeds and scales the stopwatch could not achieve — continuous monitoring, real-time scoring, instant enforcement.

Opacity of the standard. The algorithmic standard is held by the system and not fully visible to the worker, perfecting the information asymmetry Taylor sought.

Extension to knowledge work. The framework moves from physical labor to cognitive labor through tools that track digital activity the way stopwatches tracked physical motion.

The AI amplification. When knowledge workers adopt AI tools, algorithmic management converts the capability gain into intensification rather than amplification — more output expected, rest colonized, judgment subordinated to measurable activity.

The alternative available. Algorithmic management is a choice, not a requirement; organizations can deploy AI with different surrounding infrastructure, cultivating judgment rather than enforcing output.

Debates & Critiques

Defenders of algorithmic management argue that surveillance and measurement improve productivity and enable meritocratic allocation of work. Critics note that the frameworks' gains come at costs the metrics cannot capture — the erosion of judgment, the degradation of morale, the systematic miscounting of value in knowledge work where the most important activity is invisible to activity-tracking tools. The AI transition forces the question: will organizations deploy the amplifier through the framework of the stopwatch, or will they rebuild their measurement systems to capture the judgment the stopwatch was never designed to see?