The Outsourced State

When Governments Stop Governing and Start Licensing Power

Constitutions were written for institutions that could be named. Ministries. Courts. Police forces. Legislatures. Their powers, limitations and responsibilities were deliberately assigned because democratic systems assume a simple principle: public authority must remain visible. It is easier to challenge a government than a process, easier to scrutinize an elected official than a software architecture. Yet many of the systems shaping public life today occupy a space that constitutional law never anticipated. They are neither public institutions nor purely private enterprises. They are technological intermediaries through which governments increasingly exercise essential functions. The transition has unfolded without constitutional amendments or political upheaval. It has arrived through procurement contracts, cloud infrastructure and software licenses—documents that rarely attract public attention, but whose institutional consequences may prove more enduring than legislation itself.

Across Latin America’s largest cities, this evolution is no longer theoretical. Artificial intelligence assists police investigations, biometric platforms authenticate access to public services, predictive systems help allocate security resources, while integrated urban command centers promise faster and more coordinated responses to emergencies. None of these developments is inherently problematic. States have always adopted new technologies to improve administration. The legal question begins elsewhere. At what point does assistance become dependency? There is a meaningful difference between using technology as an instrument of government and allowing technological infrastructure to become a precondition for governing. That distinction rarely appears in policy debates because efficiency is easier to measure than institutional displacement. Governments often become more capable while simultaneously becoming less autonomous, a paradox that receives remarkably little legal attention.

The difficulty is not simply that algorithms influence decisions. Human institutions have always relied on external expertise, from accountants to engineers and forensic laboratories. The novelty lies in the location of authority. Increasingly, the informational environment within which public decisions are made is designed, maintained and updated by actors operating outside the constitutional chain of accountability. A municipality may own the cameras but not the software interpreting their images. A ministry may issue the decision but depend upon a proprietary model whose internal logic remains inaccessible. Citizens retain the formal right to contest administrative acts, yet the decisive layer of reasoning often resides inside systems protected by commercial confidentiality rather than public scrutiny. The law recognizes the decision-maker. It is far less prepared to examine the architecture that quietly shaped the decision before any official signature appeared.

This matters because sovereignty has never been exercised solely through laws. It has also depended upon institutional capacity. The ability to identify individuals, regulate movement, investigate crime, collect taxes or certify property has historically distinguished functioning states from fragile ones. Today those same capabilities increasingly rely on interconnected digital ecosystems whose ownership, maintenance and technical evolution frequently extend beyond national institutions. The consequence is subtle. Governments may appear stronger because they process more information, anticipate more risks and automate more procedures. Yet every additional layer of technological capacity can also create another layer of contractual dependency. Public authority expands operationally while becoming structurally contingent on infrastructures it does not fully control. The state becomes more efficient precisely where it becomes less self-sufficient.

This pattern is no longer confined to domestic governance. Security technologies now circulate across continents with remarkable speed, carrying with them assumptions about risk, surveillance and acceptable intervention. A predictive policing model refined in one metropolitan region may later influence border management elsewhere. Urban surveillance architectures developed to address organized crime can eventually inform migration policy or critical infrastructure protection in entirely different jurisdictions. These systems do not merely export software; they export institutional logic. Legal traditions travel far more slowly. By the time constitutional doctrine begins to address one generation of digital governance, another has already entered public administration. Law increasingly follows technology rather than framing it.

Perhaps this explains why the outsourcing of state functions rarely provokes constitutional debate. The transfer is incremental, distributed across hundreds of administrative decisions that appear practical when viewed individually. There is no singular moment at which sovereignty changes hands. There is only accumulation. A new platform here. A new database there. Another predictive model added to an existing system. None of these choices seems capable of redefining the relationship between citizens and the state. Together, however, they begin to redraw the institutional map upon which democratic accountability depends.

The question, then, is not whether governments should embrace artificial intelligence. They almost certainly must. The more difficult question is whether constitutional democracies can preserve public responsibility when essential instruments of governance are increasingly embedded within infrastructures they neither fully own nor fully understand. History often remembers the visible crises that transform political systems. It pays less attention to the administrative routines that quietly alter them. If the twenty-first century produces a new constitutional dilemma, it may not emerge from the rise of artificial intelligence itself. It may emerge from something far less dramatic: the gradual normalization of a state that continues to govern its citizens while relying, ever more deeply, on institutions that it does not entirely govern.

Rafael Santoro, Brazilian columnist at Phoenix24. Specialist in cybercrime, surveillance, and the technopolitical transformation of Latin American cities.