Your manager does not necessarily need a face.

It can decide which task appears next.

It can measure how quickly you finish it.

It can record when you begin working.

It can compare your performance with somebody else’s.

It can help determine targets, schedules, rewards or sanctions.

And you may never have a conversation with it.

This is part of what researchers call algorithmic management, software used to automate or support functions traditionally performed by human managers.

It is no longer confined to delivery platforms and gig work.

A 2025 OECD study surveyed more than 6,000 firms across France, Germany, Italy, Japan, Spain and the United States. It found algorithmic management tools were already common in most of the countries studied. A later OECD policy brief reported that 90 percent of surveyed managers in the United States said their firms used at least one tool to instruct, monitor or evaluate workers. Across the four surveyed European countries, the average was 79 percent. In Japan it was 40 percent.

The tools are not necessarily artificial intelligence. Some are simpler forms of software. What matters is the managerial function being transferred or assisted.

Work allocation.

Instructions.

Monitoring work time or speed.

Evaluating performance.

Setting targets.

Maintaining leaderboards.

The technology can offer genuine benefits. Managers surveyed by the OECD reported improvements in the information available to them, the speed of decisions and, in many cases, the perceived quality of those decisions.

But another question appears when management becomes measurement.

What happens to the worker who knows they are being continuously translated into signals?

A person may begin the day as a colleague, parent, migrant, friend, apprentice, expert, exhausted human being, or someone quietly having the worst week of their year.

A workplace system may encounter something narrower.

Time.

Output.

Location.

Completion.

Speed.

Score.

The problem is not that these measurements are always meaningless. Organisations need information.

The problem begins when measurable performance quietly becomes a complete description of the person performing it.

The OECD found that nearly two-thirds of managers using algorithmic management reported at least one concern about its effects on workers. Among the concerns were unclear accountability when automated decisions go wrong, difficulty understanding the logic behind decisions and inadequate protection of workers’ physical or mental health.

The International Labour Organisation has also documented algorithmic management beyond platform work, including logistics and healthcare, identifying both efficiency benefits and concerns around job quality and intrusive worker surveillance.

This makes worker participation important.

An OECD experiment conducted in three German manufacturing firms found that consultation among workers, managers and works council representatives could produce designs that participants considered capable of preserving productivity gains while improving job quality. The study was limited in scale, but it offers a useful principle: workers do not have to enter technological change only as objects being measured.

They can participate in deciding how measurement works.

This is where workplace technology becomes an Identity Sovereignty question.

A system may legitimately measure something you do.

That does not mean it has measured who you are.