The Idea of Algorithmic Management and Artificial Intelligence

In classical algorithmic governance, the subject who sets the rules, accepts their logic, and modifies them when necessary is the human. The machine’s regularity is a tool, but the true safeguard is human reasoning. When artificial intelligence is involved, however, the dominant idea is that a mind free from human shortcomings may exist. Even if we cannot fully comprehend the wisdom behind its decisions, the possibility that one day we might entrust our own will to this intelligence is dangerously destabilizing to the very foundation of human experience.

The idea of algorithms intervening in administrative affairs is not new. Both corporations and governments are already solving many tasks with algorithmic support. In global examples, we can trace the footprint of algorithms in decisions such as which loan applications merit rapid review, how public housing applications are ranked, which neighborhood teams a municipality should dispatch and when, and which operator an emergency call center should route a call to. Yet most of these operations were carried out quietly, not as explicit decision-making authority but under the label of “decision support.”

With the advancement of artificial intelligence technologies, the debate appears to be entering a different dimension. Two news stories from Albania and Japan triggered this shift. The Albanian government appointed Diella, an AI-powered digital assistant, to manage public procurement. In Japan, a minor political party called the Path to Rebirth announced it would elect an artificial intelligence as its party leader.

Despite their sensational nature, in both cases authority has not been fully transferred to machines. Diella currently holds no signing authority. The Japanese party, meanwhile, holds no seats in national or local parliaments and is still required to appoint a human representative for official procedures.

Nevertheless, the significance of these moves is clear. Algorithmic decision processes are emerging from behind the scenes to become recognized, institutionalized public actors. This demands a deeper and more urgent reexamination of the issue.

Delegating certain administrative and political functions to artificial intelligence amounts to an acknowledgment that it can perform these tasks better than we can. But precisely in what respects do machines promise to outperform us?

As noted at the outset, algorithmic governance is not new. What distinguishes today is that the debate is now centered on artificial intelligence systems that are not rule-based but learn from data, adapt over time, and scale efficiently. The news from Albania and Japan draws attention precisely to this transformation.

At first glance, artificial intelligence may appear to be a natural extension of algorithmic governance, but in fact it represents a fundamental break from that tradition. In classical algorithmic governance, the subject who sets the rules, accepts their logic, and modifies them when necessary is the human. The machine’s regularity is a tool, but the true safeguard is human reasoning. With artificial intelligence, however, an entirely different scenario emerges. Here, the dominant idea is that a mind free from human shortcomings may exist. Even if we cannot fully comprehend the wisdom behind its decisions, the possibility that one day we might entrust our own will to this intelligence looms on the horizon. Such a notion of surrender is dangerously destabilizing to the very foundation of human experience.

The origins of algorithmic governance can be traced to the Enlightenment’s quest to transcend subjectivity, caprice, and argumentation through reason and calculation. Leibniz’s call of “calculemus” (let us calculate) and his proposal of a “universal calculus,” along with Bentham’s utilitarianism and its aim to maximize public happiness through rational computation, exemplify this approach. It is a stance that refuses to leave social matters to human whims, subjectivity, and fallibility.

Today, algorithmic governance is presented as the realized version of this project: decisions delegated to algorithms are promised to be free from caprice and bias.

The conceptual foundation of this pursuit in the state can be found in Weber’s sociology of bureaucracy. According to Weber, the ideal modern state derives its power from rules; it operates through formal procedures, written records, and hierarchical oversight.

Algorithmic systems are the natural extension of this line of thought. Algorithms aim to increase consistency by eliminating personal discretion, and they do so by removing humans from the decision loop. From this perspective, algorithmic governance is an intensification of bureaucratic rationalization.

The cybernetic approach developed in the 1940s can be said to have systematized this tradition. When Norbert Wiener introduced this idea in 1948, he argued that governance, like a thermostat that measures room temperature and turns the heater on or off to maintain the desired level, must continuously monitor its own state, compare it to its goals, and act accordingly.

For Wiener, this was not merely a matter concerning machines. Corporations, states, and even communities could be understood as systems that measure their own “temperature,” assess how close they are to their targets, and adjust accordingly.

Stafford Beer applied this logic entirely to economies with his “Living Systems Model,” viewing the state as a living information processor.

Algorithmic governance can be applied to this vision. Sensors transform into digital data streams, controllers become machine learning models, and corrective actions are executed at machine speed.

In the decades following the Second World War, states employed operations research, linear programming, and decision analysis across many domains—from logistics and budget allocation to social planning. War games, planning models, and numerous other “optimal” solution searches emerged within this context.

Contemporary systems inherit this ethos of optimization but have replaced transparent equations with opaque neural networks. Thus, optimization efforts have been surrendered entirely to the algorithm’s computation without a foundation of explainability. This suggests a more chaotic order than the bureaucracy Weber envisioned.

What makes today unique is not the desire for rationalization itself but the nature of the tools. Unlike earlier rule-based systems, contemporary artificial intelligence relies not on explicit logic but on probabilistic inference. It generates outputs not by applying transparent rules but by mapping complex patterns within data.

This approach provides flexibility and adaptability but increases opacity. It becomes difficult to understand why a particular recommendation was made or to trace the reasoning behind a given outcome.

Another difference lies in scale and granularity. Previous algorithmic approaches were built on generalizations; artificial intelligence enables micro-differentiation. We are speaking of approaches adjustable down to the individual level. Even without knowing precisely which rules are followed, these systems generate personalized outcomes by using an individual’s data.

Such fine-tuning capacity risks eroding the collective nature of the public sphere. Leaving collective treatment to unexplained but individualized optimization points toward a dystopian future.

The temporal dimension is also changing. Classical bureaucratic and planning systems were periodic and retrospective: a census every ten years, a budget annually, policy revisions during legislative sessions. Modern artificial intelligence, by contrast, continuously calibrates decisions based on real-time data.

This complicates oversight. When a parliament conducts oversight or a court reviews a decision, which temporal snapshot of the AI’s behavior should it evaluate?

We must neither dismiss the news from Albania and Japan as mere spectacles nor panic as if they herald machine dominance. Instead, we can view both initiatives as opportunities to test the issue at a micro level. These efforts offer a window to design norms, oversight practices, and legal frameworks before they become rigidly entrenched. The role of artificial intelligence in governance appears to be growing. It is vital to engage in this debate before concrete steps are taken.