Software now assigns shifts, scores productivity and routes work, and the body that polices collective labor law has noticed. The NLRB's general counsel issued a memo in October 2022 telling regional offices to scrutinize algorithmic management tools under existing labor law, from surveillance of organizing to automated discipline (NLRB GC memo 23-02, 2022). The legal framework did not change; the technology did.
This analysis maps what algorithmic management actually does on shop floors and in offices, and where it collides with rules written before it existed.
What is algorithmic management, concretely?
It is the delegation of managerial functions to systems: automatic scheduling from demand forecasts, real-time productivity scoring, automated task assignment, and progressive discipline triggered by metrics. In warehousing it is scan-rate tracking; in delivery it is route and acceptance scoring; in call centers it is minute-by-minute adherence monitoring; in offices it is activity analytics on email and calendars (NLRB GC memo 23-02, 2022, cataloguing these practices).
The defining feature is not the data but the decision. When a system, not a supervisor, initiates the discipline or the assignment, the paper trail changes shape — and so does the ability of a worker to contest it.
Why does the labor board care about software?
Because the National Labor Relations Act already covers the conduct the tools enable. The 2022 general counsel memo directed regions to analyze electronic surveillance and algorithmic tools against settled doctrine: surveillance that tends to restrain organizing, rules barring workers from discussing metrics, and discipline imposed without the human discretion that bargaining agreements assume (NLRB GC memo 23-02, 2022). General counsel memos charge and frame cases; the board itself decides, and this article reports the memo as a charging posture, not as settled board law.
Reuters' coverage of workplace monitoring and organizing disputes through this period documented employers deploying monitoring systems and unions filing charges over them, with outcomes pending in multiple matters (Reuters, 2022–2023). The docket, not the memo, is where the law is being made.
Does the law require disclosure of monitoring?
Federal labor law does not, in general terms; state notice laws increasingly do. New York's electronic monitoring notice act, effective in 2022, requires private employers to give written notice of electronic monitoring of employees at hiring and to post a notice (New York law, 2022). Connecticut and Delaware have had similar notice statutes for years. Notice, however, is not consent and not bargaining.
The gap matters analytically. An employer can comply with every state notice law and still face a labor board theory that the monitoring itself, in a covered workplace, interfered with protected activity. The statutes and the act regulate different harms.
What changes when a union contract exists?
Bargaining converts the algorithm into a mandatory subject. Deploying or materially changing systems that set work pace, assign discipline or measure performance touches wages, hours and terms of employment, and unions have grieved metric-based discipline as discipline without just cause (arbitration practice; NLRB GC memo 23-02, 2022). Nonunion workplaces have no equivalent lever, which is why the same software produces different disputes in different buildings.
One reading of the evidence this publication is willing to own: the collective bargaining agreement is currently the only institution that can open the black box as a matter of right, because information rights attach to the contract rather than to the worker.
The surveillance half of the memo has its own legal anatomy. The general counsel flagged camera placement, audio recording and monitoring of protected concerted activity as facts that can support a theory of interference, and directed regions to request the relevant system documentation early in investigations (NLRB GC memo 23-02, 2022). The practical effect, whatever a board later decides, is evidentiary: employers defending monitoring decisions can now expect the systems themselves to be document requests.
The productivity-metrics half raises a doctrine most coverage misses: the duty to furnish information. Unions exercising representation rights can demand the algorithms' logic where it bears on bargaining or grievance handling, and employers have resisted on trade-secret grounds, leaving boards to balance the two interests case by case (representation-case information duties, long-standing). An algorithm that sets piece rates or triggers discipline is hard to distinguish, functionally, from the wage and discipline records unions have always been able to inspect.
The office case is where the framework is least tested. White-collar analytics measure output indirectly — activity signals, response times, collaboration graphs — and the act's coverage of professional employees is narrower and more fact-bound than its coverage of warehouses. Whether the memo's theory reaches a salaried office workforce is precisely the question no decided case has answered, which is why this analysis treats the office frontier as open ground rather than settled doctrine.
The comparative context sharpens the stakes. The European Union's platform work directive and several national algorithms-at-work proposals start from transparency duties — workers told when algorithms manage them, with human review of consequential decisions (EU-level legislative activity, 2024). The United States has no federal analogue, which leaves the NLRB memo and scattered state notice statutes as the whole domestic framework. An analyst comparing the two regimes would conclude the American system regulates algorithmic management through a labor-relations statute written in 1935, applied tool by tool, case by case — a structural fact, not a complaint.
What is established, and what is open?
Established: the general counsel's charging posture treats algorithmic management as analyzable under the act (NLRB, 2022); several states require monitoring notice (New York, 2022); covered contract workplaces can bargain over the systems. Open: whether the board adopts the general counsel's approach in a decided case, and whether Congress or state legislatures move from notice to substantive limits.
The pace question — who sets the rhythm of work — has always been the core of labor law. Algorithmic management moves the setting of rhythm from the floor supervisor to the forecast model. The law has noticed. Whether it can keep up is the open file.
Related: Workplace surveillance and the law · Reading employer health data · more in the workplace.
For more context, read Workplace surveillance and the law.
For more context, read Shift work and the body clock.
For more context, read Why hospital staffing falls short.