Employment: beyond bias audits to algorithmic management
Lesson 2 of 5 in Sector by Sector: Health, Employment, Finance, Education, and Vehicles.
You have already met hiring-AI bias law — NYC Local Law 144’s audit regime, the EEOC’s iTutorGroup settlement, Colorado’s consequential-decision duties (see the US state-laws module; we will not re-teach them). This lesson covers what those regimes miss: AI does not just decide who gets hired. Increasingly it is the manager — assigning shifts, setting piece rates, scoring productivity, and terminating.
The category is algorithmic management, and its emblem is the platform economy. Uber and Lyft drivers are dispatched, priced, and deactivated by models; Amazon warehouse workers are paced against algorithmic rate targets; Amazon Flex delivery drivers were, for years, terminated by automated email with no human decision-maker to appeal to. The governance questions differ from hiring bias: opacity of the rules of work (what drops your score?), inability to contest machine decisions, surveillance intensity, and the collective dimension — algorithmic pay-setting can individualise wages in ways that dissolve bargaining power.
The EU answered with the world’s first dedicated statute: the Platform Work Directive (in force December 2024, member-state transposition due December 2026). Its algorithmic-management chapter — which applies to platform workers regardless of employment status — prohibits automated systems from processing certain data at all (emotional or psychological state, private conversations, off-duty activity, data predicting union activity or exercise of rights), requires transparency about the automated systems that take or support decisions, mandates human oversight with qualified reviewers, and gives workers a right to a human review of significant decisions — account suspension or termination cannot rest on an algorithm alone. It also creates a rebuttable presumption of employment where facts indicate control, attacking misclassification at its root.
EU
Platform Work Directive for platforms (transposition by Dec 2026); AI Act for everyone else — employment is Annex III high-risk (recruitment, task allocation, monitoring, evaluation, termination), emotion recognition at work is prohibited (Art 5(1)(f)), and deployers must inform workers and their representatives before putting a high-risk workplace system into use (Art 26(7)).
Underneath both: GDPR Art 22 rights against solely automated significant decisions, and Art 88 letting member states set stricter workplace rules.
US
No federal algorithmic-management statute. The fragments: NLRB scrutiny of surveillance that chills organising; a (since-rescinded) White House push on worker surveillance; California’s “no robo bosses”-style bills and 2025 ADS regulations under FEHA reaching employment decisions; state wiretap/monitoring-notice laws (Connecticut, Delaware, New York for electronic monitoring).
The action is litigation and agencies: EEOC (disability accommodation vs. automated assessments), FTC (worker surveillance as unfair practice theory), and wage-hour suits over algorithmic pay.
Co-determination
In Germany and much of northern Europe, the oldest tool works surprisingly well: works-council co-determination. Under §87 of the German Works Constitution Act, introducing technical systems capable of monitoring performance requires works-council agreement — which makes every algorithmic-management rollout a negotiated one, with access to the system’s logic as the price of consent. ILO tripartite discussions are pushing this model globally; the 2025 ILO standard-setting discussions on platform work aim at a binding convention.
Key terms: algorithmic management, platform work directive, works council, presumption of employment, article 22
Interactive checkpoint quiz (2 questions) — open this page in a browser to take it.