FIELD GUIDE · WORK, MANAGEMENT & THE ALGORITHM

SHOULD AI BE ALLOWED TO FIRE YOU?

Software already schedules, monitors, scores, ranks, and filters workers. Termination is where a productivity tool turns into a question about power, explanation, discrimination, and appeal.

UPDATED 2026-09-22

THE PROBLEM

YOUR TERMINATION WAS GENERATED SUCCESSFULLY

The manager clicked approve. The score arrived first.

Algorithmic management is software that performs tasks managers used to do: assign work, monitor activity, evaluate performance, rank people, and recommend actions. A human may still click the final button, but the practical power can sit upstream in the data and score that shaped the decision.

01

WATCH

MEASUREDActivity becomes data.
02

SCORE

RANKEDA proxy becomes an evaluation.
03

RECOMMEND

RED FLAGThe machine shapes attention.
04

FIRE

WHO OWNS IT?A human signature does not answer the whole question.

THE FIGHT

WHO FIRED YOU: THE BOSS OR THE SYSTEM?

Employment decisions determine income, insurance, immigration status, housing stability, reputation, and careers. Automation can make management faster and more consistent, but it can also turn bad proxies into policy at scale and make it difficult for workers to know what evidence actually cost them a job.

THE MANAGEMENT CASE

DATA CAN HELP MANAGERS MAKE MORE CONSISTENT DECISIONS

Algorithmic tools can surface patterns, reduce repetitive supervision, allocate work, and support faster decisions across large organizations. OECD survey data shows managers who use these systems often report more information and faster decision-making.

A dashboard can reveal what one manager cannot personally observe.

THE WORKER CASE

HIGH-IMPACT DECISIONS NEED EXPLANATION AND ACCOUNTABILITY

Automated evaluation can rely on incomplete proxies, reproduce historical patterns, or penalize circumstances the system does not understand. Employment law still applies, and current litigation is testing who is responsible when AI-driven tools allegedly produce discriminatory outcomes.

A spreadsheet does not become fair because nobody remembers who wrote the formula.

THE WEIRD SHIT

THE BOSS NOW HAS A DASHBOARD FOR YOUR EXISTENCE

ALGORITHMIC MANAGEMENT IS ALREADY WIDESPREAD

A 2025 OECD employer survey across six countries found widespread adoption of software that instructs, monitors, or evaluates workers, with especially high adoption reported in the United States.

This is not only a future-of-work scenario.

MANAGERS THEMSELVES REPORT ACCOUNTABILITY CONCERNS

In the OECD survey, nearly two-thirds of managers using algorithmic management reported concerns about worker impacts; unclear accountability and inability to follow the logic of decisions were among the leading issues.

The people deploying the tools also see governance problems.

A 2026 META LAWSUIT DIRECTLY CHALLENGES AI'S ROLE IN LAYOFFS

Twenty-six former Meta employees alleged that AI-powered systems disproportionately targeted workers with disabilities or medical leave in a mass layoff. Meta disputed the claims and said humans made the decisions; the merits remain contested.

The case illustrates the exact dispute: when is an AI-influenced termination legally and practically a human decision?

AI EMPLOYMENT SCREENING IS ALREADY UNDER MAJOR LEGAL TEST

Workers in Mobley v. Workday allege AI-based screening produced discriminatory rejections. Workday denies wrongdoing, and the litigation is testing how liability should be allocated between software vendors and employers.

Automated employment decisions do not sit outside ordinary discrimination law.

THE PEOPLE WITH A STAKE

WHO GETS TO APPEAL A NUMBER?

WORKERS

They bear the direct consequence of an incorrect score, opaque evaluation, or automated recommendation that shapes access to work.

MANAGERS

They gain scale and information but may become dependent on tools they cannot fully interrogate.

EMPLOYERS

They capture efficiency and remain exposed to legal and reputational consequences when automated decisions violate worker rights.

SOFTWARE VENDORS

Their models can become de facto decision infrastructure while courts and regulators debate how much responsibility travels with the tool.

THE UNANSWERED QUESTION

WHAT MUST A HUMAN ACTUALLY DO BEFORE A MACHINE-INFLUENCED FIRING?

TAKE THIS TO DINNER: The machine can produce a score. Someone still has to own the firing.

RECEIPTS

The lawsuit claims are contested. The management software is not imaginary.

  1. How widespread is algorithmic management in workplaces?OECD
  2. Meta used AI to target workers with medical conditions for layoffs, lawsuit claimsReuters
  3. These workers say an AI hiring platform discriminated against them. But who should pay?Reuters