When AI becomes your boss: The limits of algorithmic management

This post was originally published on this site.

Algorithms are increasingly deciding when Europeans work, how their performance is measured and, in some cases, how much they earn. Tasks once reserved for human managers are gradually being handed over to computer programs. And as artificial intelligence and algorithmic management spread beyond platforms such as Uber, Deliveroo and Glovo, EU rules on data protection, transparency and workers’ rights are struggling to keep pace.

A 2026 EU survey found that 42% of European small and medium-sized enterprises and mid-cap companies already use AI and algorithmic management, with another 21% planning to do so. A separate EU study found that workers subject to algorithmic management generally report higher levels of job strain than those who are not, particularly when it comes to automated scheduling and performance monitoring.

“Overall, extensive research shows that the use of AI and AM tends to have more benefits for those who manage and design those systems than those who are below them,” said Davide Monaco, research fellow on the future of work at the Centre for European Policy Studies (CEPS).

The result, Monaco explained, is a growing perception among workers that they are working for machines rather than with them. “They’re dehumanising work and that has negative impacts on the relationship between employers and employees.”

Recent disputes illustrate the tensions. Uber drivers in several European countries have launched legal action over the company’s algorithmic pay-setting systems. Amazon has faced scrutiny over performance monitoring in its warehouses, while delivery couriers in Britain have demanded greater transparency from food delivery platforms over how decisions on pay and work assignments are made.

As labor unrest spreads across parts of Europe, the growing use of automated management systems is also raising questions about accountability when algorithms determine pay, terminate employment or contribute to unsafe working conditions.

The European Commission is expected to address some of these concerns in its Quality Jobs Act proposal, due December 2. Business groups oppose another layer of regulation, while some lawmakers and researchers warn that existing safeguards are insufficient to prevent the balance of power from shifting further toward employers.

“If workers cannot shape these systems whenever they are adopted and implemented and if collective voice is not strengthened, people are going to find other ways to express their discontent,” Monaco said.

New forms of work, new rules?

Europe is hardly short of digital regulation. Its strict data protection rules, landmark AI Act and legislation on platform work already place limits on how employers can use technology. But taken together, they leave significant gaps in the protection of workers — gaps that could widen as Brussels moves to simplify its digital rulebook.

A recent report by independent researchers presented to the European Parliament’s Employment Committee warned that shortcomings in the General Data Protection Regulation (GDPR), AI Act and Platform Work Directive could be exacerbated by the Commission’s proposed digital omnibus to simplify the bloc’s tech rulebook.

An upcoming study by the European Trade Union Institute (ETUI), seen by The Parliament, raises further concerns. Among the proposed changes are amendments to GDPR Article 12(5), which could make it easier for companies to refuse requests for personal data, and Article 22, which could broaden the circumstances in which employers justify decisions made entirely by algorithms. A proposed provision on AI-related data processing would also explicitly recognize legitimate interests as a legal basis for developing or operating AI systems, subject to safeguards.

The researchers tested the changes against four workplace cases involving Amazon, H&M, Uber and Ola, and Glovo. They found that the original legal outcomes would remain broadly attainable, although protections in the Uber and Ola cases would be weakened. The greater risk, they argued, is that similar violations could become harder to uncover. With less access to information, workers may struggle to challenge automated decisions, while data protection authorities could find it harder to investigate abuses.

Such concerns are closely tied to the limits of human oversight, a concept that existing EU laws interpret differently depending on their regulatory purpose.

“Human oversight from a data protection perspective is not the same as from an algorithm management perspective,” said Aida Ponce del Castillo, an ETUI senior researcher and co-author of the study.

For oversight to mean anything, she argued, workers must know when an algorithm is being used and what it does, and be able to challenge decisions when something goes wrong. Yet many data protection rights depend on information employees may not have. Workers cannot contest monitoring systems they do not know exist, much less automated decisions whose underlying data and criteria remain opaque.

Existing privacy laws have nevertheless provided a means of challenging some of these practices. In France, the data protection regulator fined Amazon €32 million in 2024 for excessively intrusive monitoring of warehouse employees. Similar GDPR enforcement actions against Uber and Glovo have targeted automated decisions affecting drivers and couriers.

For Alex Saliba, a Maltese lawmaker and vice president of the Socialists and Democrats group for social Europe, the potential productivity gains from these technologies should not come at workers’ expense.

“AI should work to increase productivity, but productivity should not be increased by treating workers as robots,” he said.

Workers themselves appear less opposed to the technology than to its potential misuse. A 2025 Eurobarometer found that 62% of respondents viewed robots and AI positively at work, while 84% believed their use required careful management to protect privacy and ensure transparency.

“This is the evolution of how we are working. We cannot stop it,” Saliba said. “But at the same time, we need to understand how to use it best.”

For him, that starts with giving workers and their representatives a greater say in how workplace technologies are designed and deployed.

The next regulatory battle

The Commission’s response is taking shape in the Quality Jobs Act, an initiative first outlined in 2024, when Ursula von der Leyen tasked Executive Vice President Roxana Mînzatu with examining the impact of digitalization on work. Its scope extends well beyond algorithmic management to include occupational health and safety, subcontracting, labor enforcement and the green and digital transitions.

Whether the proposal will impose binding obligations remains unclear. But the prospect of new legislation has already exposed a familiar divide between employers and trade unions.

BusinessEurope, the European employers’ association, sees little reason to add another layer of regulation. In its latest position paper, it argued that “the only appropriate way forward on AM and AI” is supporting member states in enforcing existing EU rules rather than creating new ones. More legislation, it warned, would raise compliance costs, slow AI adoption and leave European businesses further behind their global competitors.

Trade unions, researchers and some lawmakers see the opposite risk: that existing protections are ill-suited to the technologies increasingly governing workers’ lives.

Even among those calling for stronger safeguards, however, the Commission’s approach has raised concerns. Saliba fears that bundling so many labor issues into a single initiative could force lawmakers to negotiate competing priorities rather than address each problem on its merits.

“My biggest worry is: are we going to end up negotiating on what should be getting top priority under this bag of different rights where one is not directly linked to the other?” he said.

Andersson, meanwhile, argues that a single set of rules tailored to the labor market would benefit companies as well as workers. Without common EU standards, she warns, employees could face different levels of protection across the bloc, while businesses could gain a competitive advantage by operating in countries with weaker safeguards, distorting competition within the single market.

“This will have to be addressed sooner or later,” Andersson said. “In labor market policy, problems often emerge in one sector and, if left unaddressed, spread elsewhere.”

Follow the developments shaping Europe

Subscribe to The Continental to receive The Parliament’s latest reporting, analysis and opinion to your inbox, delivered every weekday morning. Sign up for free here.

Leave a Reply

Your email address will not be published. Required fields are marked *