Banking without humans: A case the courts cannot yet hear

___________________

OPINION

Dr Theodora Twongyirwe Mondo

The more I read about emerging technologies, the more conflicted I become. On one hand, they carry the promise of improved service delivery, new efficiencies, and promises of new professions for the coming generation.

On the other hand, they carry consequences we have not fully named. In my youth, I watched science-fiction worlds like I Robot, where humanoid machines walked among us as security officers and decision-makers; or Blade Runner, where artificial beings blurred the moral line between human and machine; and later Westworld, where robots did not only serve but eventually demanded rights. I never imagined I would live long enough to see fragments of those imagined futures leak into reality — yet here we are, and they have arrived without knocking.

In the last decade alone, traditional bank branches have been shutting down at a historic pace. The United States has closed more than 10,000 branches, the United Kingdom has lost over 5,000, and Asia’s fastest-growing banks now operate with zero physical branches. Africa is not exempt from this migration: more than 70% of new financial accounts on the continent are now opened digitally through mobile money and fintech platforms, not inside banking halls.

Banks are no longer built out of granite and glass — they are moving into algorithms, APIs and AI-driven onboarding systems. As banks embrace AI-based onboarding, Uganda must confront the legal vacuum around machine-made financial decisions. This is not academic speculation.

Uganda has witnessed an exponential rise in fintech licences, AI-driven fraud systems, automated credit scoring and instant digital onboarding in just the last five years. Banks are quietly turning into software companies with banking licenses. Much of the customer journey, from risk assessment to approvals, is already mediated by algorithms. At this pace, even your bank loan may soon be approved before you finish complaining about it.

Certain thoughts, therefore, no longer belong to fiction. There will come a day when a robot can hire or fire a human being. There will come a day when an AI agent can incorporate a company, open a bank account and complete compliance checks without a single human signature. A branch manager may soon be replaced by code that never asks you to “come back tomorrow.” All of this is technologically plausible. But is the law prepared for such a world?

Uganda’s legal architecture, from the Bank of Uganda’s compliance frameworks to the National Payment Systems Act, the Companies Act, the Data Protection and Privacy Act, and even our theory of legal evidence, all assume that the decision-maker is a human being who can be summoned, examined, held to account or punished. AI does not fit that template. A machine cannot take an oath. A machine cannot hold criminal intent. A machine cannot pay damages or serve a sentence. Yet we are steadily migrating consequential human functions into its hands, as though the judges, not the machines, will be the ones to adjust.

The risk is not that AI will misbehave; all systems misbehave. The risk is that when AI makes a consequential decision, our legal system may have no defined subject of liability. If an AI onboarding tool fraudulently opens an account, who is liable: the bank, the AI vendor, or no one at all? If AI silently excludes certain groups from financial access, is that discrimination in Ugandan law or just “a system error”? If Ugandan personal data is processed offshore, which jurisdiction governs the harm? If Parliament waits for the first disaster to legislate, AI will already have grandchildren by then.

These are precisely the kinds of frontier questions that should already be on the desk of the Uganda Law Reform Commission. It must now interrogate AI-mediated decision-making in finance, employment and commerce before evidence outruns doctrine and machines outrun statutes.

If I were a young Ugandan today with dreams to pursue the legal profession, this is exactly where I would plant my intellectual career, at the intersection of AI and the law. The first litigation, the first constitutional challenges, and the first regulatory hearings of this AI era are not decades away — they are calendar-close.

Technology will not wait for us to become ready. The machine is already taking its seat. The question is whether the law will arrive before or after the verdict is delivered, perhaps by the machine itself.

The writer is a Senior Lecturer at Mbarara University of Science and Technology (www.must.ac.ug). Email tmtwongyirwe@must.ac.ug