Algorithms don dey control plenty tins for governance for long. Dem dey decide which job advert go reach which people, which tax return dem go check, which welfare case dem go handle first, and even how police go plan dia patrol waka.

Most of dis tins dey happen for back, under wetin dem dey call “decision support,” no be say dem dey make di decision directly.

Wetin make wetin dey happen now for Albania and Japan different be say di systems no dey hide again. For Albania, di government don give dia digital assistant, Diella, di work to manage procurement processes. For Japan, one small political party wey dem dey call Path to Rebirth don talk say dem go make AI dia leader.

But dis no mean say machines don take over everything. Diella still dey work under human supervision, and di Japanese party no get seat for national parliament, so dem still need human being to represent dem officially.

Even like dat, dis move dey important. E don carry algorithmic decision-making from back to di front, where everybody fit see am. E mean say di way wey algorithms dey govern now don turn to public matter wey people go need discuss well.

Algorithmic governance no be new tin. For decades, government and companies don dey use scoring formulas, risk models, and decision trees to control tins. But di kain AI wey we dey see now dey learn from data, dey change over time, and fit operate for large scale.

Dis kain system no just dey follow fixed rules. E dey generate patterns, rank options, and sometimes suggest actions wey di designers no even plan before. Dis make dem powerful, but e also make dem hard to understand.

Di public appointment of AI systems for Albania and Japan don show how tins dey change. Dis one no be just about using AI, but about how di system dey evolve and how we fit still make sure say e dey follow democratic principles.

From di time of Enlightenment thinkers like Leibniz and Condorcet, people don dey dream say calculation fit replace argument. Leibniz even talk about one “universal calculus” wey go help settle disagreement. Jeremy Bentham later carry di idea go utilitarian policy, say di goal of governance na to use rational calculation to make everybody happy.

Today, algorithmic governance dey try bring dis dream to life. E dey promise decisions wey no get bias, wey dey consistent like computer function call. But e also dey bring di fear of too much control.

Max Weber, di sociologist, describe di modern state as one wey dey follow rules, get formal procedure, and dey keep record. Algorithmic systems dey continue dis idea, but dem dey make di “iron cage” wey Weber talk about even tighter.

Later, cybernetics for di 1940s come see governance as feedback control. Norbert Wiener talk say biological, mechanical, or social systems fit dey regulate demself by sensing dia state, comparing am to goal, and correcting am. Stafford Beer for di 1970s even apply dis idea to whole economies.

Algorithmic governance dey use dis same idea. Sensors don turn to digital data, controllers don turn to machine learning models, and corrections dey happen fast like machine speed.

For di post-war era, governments dey use operations research and decision analysis to plan logistics, budgets, and social planning. But di methods before dey clear, so policymakers fit see how di numbers dey work. Now, modern AI systems dey use neural networks wey no dey easy to explain.

Di digitisation wave for di 1990s and 2000s focus on efficiency, like online portals and electronic filing. But algorithmic governance don move from just keeping record to actively steering decisions.

Di new AI tools no dey follow fixed rules like before. Dem dey use statistical inference, wey mean say dem dey map complex data correlations. Dis dey make dem flexible, but e also make dem hard to understand or explain.

Di scale and granularity of modern AI systems dey different. Before, administrative systems dey apply general rules to everybody. Now, machine learning fit target individuals or small groups, wey fit reduce waste but also raise questions about fairness.

AI systems fit also dey work in real-time, dey adjust decisions as new data dey come. Dis one fit make governance dynamic, but e go hard to monitor or audit because tins dey change steady.

Governance by AI dey shift di role of public officials. Dem no just dey govern people again, dem dey govern models too. Di question now na how to make sure say di AI dey learn in line with democratic values.

Di experiments for Albania and Japan dey show di early stage of algorithmic governance. Instead of to fear or dismiss dem, we fit use dem to design di rules, audit practices, and legal frameworks wey go guide dis new way of governance.