Chapter 8
The Relocation of Agency
Traditional forms of power were often easier to identify.
A ruler issued an order. A manager made a decision. A judge delivered a judgment. A person could usually point to the individual or group responsible for an action and ask what reasons justified it.
Modern systems are different.
Power increasingly operates through networks of decisions distributed across organizations, procedures, technologies, and information systems. Outcomes emerge from interactions among many actors rather than from a single visible source of authority.
This transformation has many benefits. Complex societies could not function if every decision required direct personal deliberation by everyone affected. Modern medicine, transportation, communication, finance, and public administration all depend on systems that coordinate more information than any individual could process.
But complexity creates a new moral problem.
When agency is distributed, responsibility can become difficult to locate.
A decision may be the result of a policy written by one group, implemented by another, interpreted by a third, and automated by a fourth. Each participant may have only partial influence over the final outcome. When something goes wrong, responsibility can appear to disappear into the structure itself.
The system did it.
But systems do not make decisions in the same way individuals do. They operate through human choices embedded in procedures and tools. The challenge is not that agency has vanished. It is that agency has moved.
This relocation changes the nature of accountability.
A person can be held responsible for a direct action. But modern power often works through design choices: what information is collected, what options are available, what incentives are created, what outcomes are rewarded, and what forms of failure are considered acceptable.
The person who chooses the architecture of a system may have more moral responsibility than the person who follows the system’s immediate instructions.
This is why automation raises questions that go beyond efficiency.
When a decision is automated, the obvious temptation is to treat the system as neutral. The machine appears to remove human judgment. But every automated system contains judgments about what matters, what counts as relevant information, and what tradeoffs should be made.
Automation does not eliminate values.
It embeds them.
The moral question is therefore not simply whether a system produces accurate outcomes. It is whether the reasons built into the system can be defended to those affected by it.
A system can be technically sophisticated and morally inadequate. It can perform a task efficiently while embodying assumptions that would not survive scrutiny if stated openly.
This is another form of agency relocation. Decisions that once appeared as explicit choices become hidden within structures.
That hiddenness matters because justification requires visibility. People cannot meaningfully evaluate a rule they cannot understand. They cannot contest a decision if responsibility has been dispersed so widely that no one can explain why it occurred.
This creates what might be called a recognition problem.
Power becomes easier to exercise when the people affected by it cannot identify where decisions come from, who is responsible, or what reasons support them.
The solution is not to eliminate complexity. Complexity is unavoidable.
The solution is to preserve pathways through which reasons can travel.
A person affected by a system must have some way to understand the system’s decisions, challenge them, and participate in the process by which they are justified. Otherwise, the system may become effective at producing outcomes while losing the ability to explain why those outcomes should be accepted.
Modern power therefore depends on a difficult balance. It must use systems capable of handling complexity without allowing complexity to become a shield against accountability.
The more power moves into structures, the more important it becomes to ask where the agency went.