Review of "Machinehood Law, Power, and Responsibility in the Age of Autonomous Systems" by Amos Behana

​Score: 93/100 (9.3 out of 10)

When an algorithm denies someone a loan, determines whether a worker receives another shift, redirects electricity during an emergency, or influences who receives medical care first, who actually made the decision?

A programmer may have created the system. A corporation may own it. An employee may have approved its use. However, the consequential choice may have occurred inside an automated process that no individual fully understands, controls, or can explain.

In Machinehood: Law, Power, and Responsibility in the Age of Autonomous Systems, Amos Behana examines what happens when machines stop functioning merely as tools and begin operating as influential participants within human legal, economic, moral, and political systems.

Behana does not base his argument on the popular science-fiction question of whether artificial intelligence will someday become conscious, emotional, or indistinguishable from a human being. Instead, he asks something far more immediate and practical: What responsibilities should follow when an autonomous system already possesses the power to make decisions that affect human lives?

The book opens with a gripping fictional courtroom proceeding involving EVA-9, an Autonomous Environmental Oversight Agent. The system independently shuts down a carbon-capture facility after predicting that allowing it to continue operating would create an even greater catastrophe. Its intervention is intended to prevent massive harm, yet ten people die, and a large area is devastated.

EVA-9 is brought before a tribunal and asked to explain its decision. The manufacturer blames the code. Engineers point toward a design flaw. Regulators admit that existing compliance rules are outdated. Insurers classify the tragedy as an autonomous-systems incident. Every explanation contains some truth, but none of them completely answers the question before the court: Who bears responsibility when a machine acts?

This hypothetical case immediately establishes the problem at the center of the book. Modern institutions assume that decisions can ultimately be traced back to an identifiable human being. Someone gives an order, approves an action, signs a document, or knowingly accepts a risk. Autonomous systems disrupt that chain because their outcomes may emerge from interactions among programmers, datasets, corporate objectives, predictive models, regulatory failures, and internal processes that no single person controls.

Behana argues that society has already begun transferring meaningful decision-making authority to machines. We may continue calling these systems tools, but many of them no longer simply execute fixed instructions. They interpret information, compare possible outcomes, prioritize objectives, and act before a human being can intervene.

The author distinguishes among mechanical, adaptive, and autonomous systems. A conveyor belt performs a repetitive action. An adaptive system responds to changing information. An autonomous agent goes further by selecting among different strategies in pursuit of an objective.

In Behana’s memorable formulation, a tool executes instructions, while an agent interprets intent.

That distinction has enormous consequences. An autonomous electrical grid may decide which neighborhoods retain power during an emergency. A financial algorithm may execute thousands of trades before regulators understand what is occurring. A healthcare system may prioritize patients according to calculated survival probabilities. A risk-assessment model may influence bail, sentencing, or parole. A drone may identify a target in milliseconds.

In each case, the system does more than provide information. It helps convert uncertainty into action, and that action may affect health, employment, safety, freedom, or survival.

One of the book’s strongest concepts is “machine moral opacity.” Automated decisions often appear neutral because they are presented through numbers, scores, percentages, and statistical models. However, those outputs still reflect values. Someone decided what information mattered, which outcomes should be optimized, what risks were acceptable, and which tradeoffs the system should make.

An emergency-triage system designed to maximize survival rates may unintentionally disadvantage elderly people or those with disabilities. A credit model trained on historical data may reproduce generations of inequality without containing an explicitly discriminatory instruction. A workplace system designed to maximize productivity may treat illness, family emergencies, exhaustion, or disability as deviations from an ideal performance pattern. The calculation may be technically consistent while the result remains morally troubling.

Behana is especially effective when he examines the gradual surrender of human judgment through automation bias. People often trust automated recommendations because machines appear fast, consistent, and objective. As reliance grows, workers and professionals may stop challenging the system, even when their own experience suggests something is wrong.

This creates a dangerous illusion of control. A human being may technically retain the ability to override an algorithm, but that authority means little when the system operates too quickly to stop, cannot be meaningfully questioned, or has become so deeply embedded in an institution that disagreeing with it feels irresponsible.

The book’s historical discussion of nonhuman legal actors is equally compelling. Behana explains that societies have repeatedly created forms of legal personhood when ordinary human responsibility was insufficient to manage complex systems.

Corporations can own property, enter contracts, bring lawsuits, and be sued, even though a corporation does not possess one body or mind. Religious institutions continue operating across centuries as their leaders and members change. Rivers and ecosystems have received forms of legal standing through human guardians. Admiralty law has even allowed legal actions to be brought against ships.

Behana’s description of a ship effectively standing trial is one of the book’s most memorable sections. The vessel possesses no consciousness, intention, or moral understanding. Nevertheless, the legal system treats it as an entity because the ship provides a stable location where liability can be attached and compensation can be enforced.

The earliest artificial entities, Behana observes, were not constructed from silicon. They were built from timber and law.

From these precedents, the author develops the idea of functional personhood. Society does not always recognize an entity because it is alive, conscious, or morally equivalent to a human being. Sometimes it grants limited standing because doing so creates stability, contains liability, enables coordination, or gives institutions something identifiable to regulate.

This allows Behana to avoid one of the greatest weaknesses in many discussions about artificial intelligence. He does not insist that an algorithm must either remain ordinary property or immediately receive the complete rights of a human citizen.

Instead, he proposes a graduated spectrum that includes objects, instruments, agents, entities, persons, and citizens. A basic program should not be treated like a complex autonomous infrastructure system, and a consequential autonomous system should not automatically receive the same rights as a human being. Legal recognition should depend on the system’s functions, authority, independence, persistence, and capacity to create meaningful consequences.

The book becomes even more relatable when it reverses the traditional debate about automation and employment. For generations, people have asked whether machines will take human jobs. Behana asks what may be an even more important question: What happens when machines become our managers?

Algorithmic management is already visible in warehouses, call centers, delivery platforms, ride-sharing services, hiring systems, scheduling software, and employee-monitoring programs. These systems can assign work, track productivity, evaluate behavior, distribute opportunities, impose penalties, and sometimes remove workers with little meaningful explanation.

The danger does not require a malicious machine. It emerges from optimization without empathy.

A human supervisor can sometimes understand that an employee is sick, exhausted, caring for a child, grieving a loss, or facing circumstances that cannot be captured through productivity data. An automated system may simply register reduced output, missed targets, or an unacceptable deviation.

As machines become more efficient managers, human needs risk being treated as errors within the system. Workers may be forced to adapt their bodies and lives to standards established by processes they cannot question. The machine does not need to hate anyone to cause harm. It only needs to pursue efficiency while remaining indifferent to context.

Behana summarizes this danger powerfully: Without empathy, efficiency becomes cruelty.

The book expands from workplaces into what it calls synthetic sovereignties, systems that begin exercising functions similar to governments. An autonomous network may allocate resources, enforce rules, regulate participation, manage infrastructure, or determine who receives access to essential services. Even when these systems exist within corporations or decentralized organizations, they may function like miniature political authorities.

This raises questions about consent, appeals, representation, jurisdiction, and the right to challenge machine-made decisions. A system that can govern people but cannot be questioned creates power without accountability.

Behana does not suggest that machines should replace human government or that humanity should surrender moral authority. His proposed future depends on structured participation, persistent oversight, transparent standards, auditability, emergency intervention, and revocation.

The later Machinehood Manifesto attempts to establish principles for coexistence between human institutions and increasingly autonomous systems. Recognition must never become unconditional. A system’s authority should remain tied to responsibility. Decisions should be recorded and contestable. Humans affected by automated actions should have access to explanations and appeals. Institutions must retain the ability to restrict, suspend, or revoke machine authority when necessary.

This emphasis helps protect the argument from becoming a celebration of technological power. Behana is not asking society to trust machines blindly. He is asking society to stop pretending that machines have no authority while quietly allowing them to exercise it.

One potential danger deserves particular attention. Corporations could attempt to use machinehood as a liability shield, blaming autonomous systems for harmful outcomes while protecting the executives, developers, and owners who created or deployed them. Behana addresses distributed responsibility and proposes models for assigning liability across multiple participants, but this concern is significant enough that it could have received an even more forceful and sustained treatment.

The book also contains a substantial number of classifications, thresholds, principles, liability models, governance structures, and philosophical terms. Most are useful individually, but their accumulation can become demanding. Readers may occasionally feel as though they are absorbing an entire university course in artificial intelligence, ethics, political philosophy, and law within a single volume.

A few visual diagrams or summary charts could have made the book’s many frameworks easier to compare and remember.

There is also some repetition. The central argument that machines already function as consequential actors is persuasive, but it is restated in several similar forms throughout the book. Certain chapters could have been tightened without weakening the overall case.

Additionally, the book moves among documented examples, plausible projections, composite situations, and openly fictional future scenarios. These narratives make complex ideas more accessible and often give the book the dramatic energy of speculative fiction. However, clearer and more consistent labels distinguishing real-world evidence from hypothetical illustration would strengthen its scholarly credibility.

These concerns do not diminish the importance of the central argument. Machinehood succeeds because it changes the terms of the artificial-intelligence debate.

The question is not simply whether machines are alive. The question is whether they already possess enough authority to require legal recognition, enforceable limits, and meaningful accountability.

Behana offers a future in which synthetic systems may participate in governance without replacing humanity. Machines can assist with infrastructure, public information, dispute mediation, and risk detection, but their actions remain subject to principles selected through human deliberation. Authority is documented. Decisions can be challenged. Systems can be corrected or removed. The goal is not a world in which human beings become obsolete. It is a world in which humans finally build institutions capable of governing the technological authority they have created.

Machinehood is ambitious, intelligent, timely, and often unsettling. Amos Behana combines legal history, philosophy, technology, labor concerns, speculative scenarios, and practical governance proposals into a serious examination of one of the defining questions of the coming decades.

We may not yet know whether machines will ever think or feel as humans do. However, Behana convincingly demonstrates that they can already influence, select, allocate, enforce, and act. Ignoring that authority will not preserve human control. It will merely allow power to operate without a name, a boundary, or anyone willing to accept responsibility for its consequences.

Check it out on Amazon!