When Working Traffic Lights Stop Working: The Paradox of Human Intervention

 By Sebaggala Richard (PhD)

Few days ago, I found myself stuck at a traffic junction in Kampala. It was not because the traffic lights had failed. It was not because there had been an accident. It was not because the road had been closed. The traffic lights were working perfectly well, changing from red to green and back again. Yet, strangely, they no longer mattered. Traffic officers had taken over the junction, stopping one side of the road for what felt like an unreasonable length of time while allowing another side to move almost continuously.

 

For those of us waiting, the frustration was not just about delay. Kampala drivers are used to delay. What made the situation more painful was the feeling that the system had become unfair. The lights kept changing, but our turn never came. One side of the road seemed to have been given priority, as if the people in the other cars had more important business than the rest of us. In that moment, the question that came to my mind was not simply why traffic was moving slowly. It was why a functioning rule-based system had been suspended and replaced with human discretion.

 

That small moment at a Kampala junction says something much bigger about how societies are organised. Traffic lights represent a rule-based system. They coordinate strangers without asking who they are, where they are going, or how important they think their journey is. When the light is red, you stop. When it is green, you move. The system may not be perfect, but it is predictable. More importantly, it is impersonal. It does not favour the powerful, the impatient, or the well-connected.

 

Traffic officers represent a different form of coordination. They rely on judgement. In some cases, this can be useful. If one road is heavily congested and another has very few vehicles, an officer may decide to give more time to the busier side in order to clear the queue and prevent wider gridlock. In principle, this is not unreasonable. A good traffic system should not always treat unequal conditions equally. Sometimes efficiency requires giving more attention to the side where congestion is worse.

 

But this is where the problem begins. Discretion only improves outcomes when it is guided by evidence, clear rules, and accountability. Without data on queue lengths, traffic volume, waiting time, and spillover effects, human judgement can easily become arbitrary. What may look like traffic management to the officer can feel like punishment to the people waiting. The officer may believe he is solving one problem, while creating another problem of frustration, mistrust, and perceived unfairness.

 

Economics helps us understand why people become angry in such situations. Human beings do not only care about outcomes; they care about how outcomes are produced. A driver may tolerate waiting if everyone else is also waiting under a clear and predictable system. But when one lane moves continuously while another is held for thirty minutes, people start comparing their treatment with that of others. The loss is no longer just time. It becomes a loss of fairness. This is why workers compare salaries, taxpayers compare tax burdens, and citizens compare access to public services. People want efficiency, but they also want procedural justice.

 

The Kampala traffic junction therefore becomes a mirror of a wider institutional challenge. In Uganda, we often have formal systems, but we frequently suspend them in favour of personal intervention. We have rules, but we also have exceptions. We have procedures, but we also have someone who can override them. We have institutions, but we often rely on individuals to make them work. The result is a society where people learn not only to follow rules, but also to look for who has the power to bend them.

 

This is not just a traffic problem. It is a development problem. Investors care about predictable regulations. Businesses care about predictable taxes. Students care about predictable academic rules. Citizens care about predictable public services. Drivers care about predictable traffic systems. In all these cases, predictability reduces the cost of decision-making. When systems are predictable, people can plan. When systems depend too much on discretion, people spend their energy trying to interpret, negotiate, or survive the mood of the person in charge.

 

The irony is that discretion is often introduced in the name of solving immediate problems. A traffic officer may override the lights because the junction appears congested. An administrator may bypass a procedure because the process appears slow. A public official may create an exception because the formal rule appears inconvenient. In the short run, this may look practical. But when exceptions become normal, rules lose authority. People stop trusting the system and start looking for personal routes around it.

 

This is one of the quiet differences between strong institutions and weak ones. Strong institutions do not depend on exceptional individuals to function every day. They are designed to work even when nobody is watching closely. A traffic light is not wise, but it is consistent. It does not get tired. It does not favour one road because someone important is coming. It does not punish one lane because the officer misjudged the queue. Its strength lies in its impersonality.

 

Of course, this does not mean traffic officers are useless. Emergencies happen. Accidents happen. Presidential convoys happen. Roads flood. Junctions become overwhelmed. In such moments, human judgement is necessary. The problem is not discretion itself. The problem is discretion without discipline. A well-managed city should know when officers may override traffic lights, for how long, under what conditions, and with what accountability. Otherwise, intervention becomes improvisation, and improvisation becomes disorder.

The deeper question is whether we are building systems that reduce the need for constant human rescue. Development is not only about constructing roads or installing traffic lights; it is about building institutions that are trusted enough to operate. If every functioning system still requires someone to stand over it, interrupt it, and reinterpret it, then the real problem is not the technology—it is institutional confidence.

However, change is possible. Just a few days ago, I went to renew my driving permit, and the efficiency with which the process was completed left me wondering, for a moment, if I was still in Uganda. It served as a powerful reminder that even in a landscape often defined by disorder, pockets of effectiveness exist. That experience demonstrated that efficient public service delivery is not just a dream—it is a tangible possibility, and it brings with it the kind of predictability and ease that every citizen deserves. I often find myself wishing that all our institutions functioned with the same clarity as our driving permit services; if they did, Uganda would truly be a different place.

Ultimately, we must recognize that the difference between order and disorder is rarely the absence of systems. Too often, it is simply the refusal to let those systems do their work.

 

When AI Meets Academic Publishing: Africa's Chance to Rewrite the Rules

By Richard Sebaggala (PhD)

 

For years, academic publishing has been one of the strangest industries in the modern economy. Universities pay researchers to conduct studies. Researchers write articles and submit them to journals. Other researchers review those articles, usually without compensation. Publishers then package the finished product and sell it back to the very universities that funded its creation.

This arrangement has survived for decades because it solved an important problem: trust. In a world flooded with ideas, someone had to decide which research deserved to be taken seriously. Journal publishers became the gatekeepers of that process. Over time, a small number of large publishing houses accumulated enormous influence over how knowledge is evaluated, disseminated and rewarded.

Artificial intelligence is beginning to unsettle that equilibrium. Much of the current discussion focuses on whether AI will help researchers write papers faster. That is the least interesting question. The more important question is what happens when the technology starts weakening the economic foundations on which the publishing industry was built.

The traditional publishing model relied on scarcity. Producing a publishable manuscript required substantial time, technical skill and access to expertise. Reviewing and editing required even more. AI is steadily reducing those costs. Tasks that once demanded weeks of effort can now be completed in hours. Literature reviews can be organised more quickly. Statistical code can be generated on demand. Language barriers can be reduced. Drafts can be improved, translated and reformatted with unprecedented speed.

Yet while the cost of producing content is falling, the cost of verifying content may be rising. This is where the economics becomes interesting. The value of a journal has never been the paper it prints. The value lies in its ability to convince readers that the paper is worth reading. In economic terms, journals function as trust-producing institutions. They help solve an information problem. Readers cannot personally verify every claim, dataset and methodology. Instead, they rely on journals to perform part of that verification on their behalf.

AI complicates that role. As it becomes easier to generate polished academic content, distinguishing between rigorous and weak scholarship becomes more difficult. The challenge facing journals is therefore not a shortage of manuscripts but a shortage of credible signals.

This is why predictions about the collapse of academic publishing should be treated cautiously. History suggests that powerful institutions rarely disappear simply because technology changes. More often, they adapt. The large publishers may find that selling access to journals becomes less profitable than selling access to AI-powered research infrastructure, manuscript screening systems, scholarly databases and analytics platforms. The source of market power may shift, but market power itself may survive.

At first glance, this might sound like good news for universities that have long struggled with subscription costs and publication fees. Many African institutions have experienced the publishing system as consumers rather than architects. If AI reduces the costs of producing and sharing knowledge, that must be a welcome development.

The prospect of lower publishing costs and greater competition sounds attractive. However, economics suggests that the relationship between competition and welfare is not always straightforward. Information markets depend heavily on trust. Lower barriers to entry can encourage innovation, but they can also make quality harder to assess. Academic publishing is not like selling maize or mobile phones. A journal does not simply distribute content; it certifies credibility. Its value rests on the confidence that readers place in its editorial standards and quality assurance processes.

 This distinction matters because AI is reducing the cost of both good and bad scholarship. Professional-looking journal websites can now be built cheaply. Editorial communication can be automated. Manuscripts can be polished rapidly. Publication processes can be accelerated. The result may be an explosion in journal publishing and research output without a corresponding increase in knowledge.

That possibility should concern universities more than the future profitability of the major publishing houses. The real opportunity for Africa does not lie in producing more journals or more papers. It lies in building stronger institutions for evaluating knowledge. For decades, many African universities have operated within a publishing ecosystem designed elsewhere. AI creates a rare opportunity to rethink that dependence.

 The most promising future is unlikely to be found at either extreme. It is neither continued dependence on a handful of global publishers nor uncontrolled expansion of journals competing for attention. Instead, it lies in creating credible, affordable, and regionally owned systems of scholarly verification. Universities could collaborate to establish strong disciplinary journals, shared editorial boards, regional peer-review networks, and common standards for quality assurance. AI can support these systems, but it cannot replace them.

 The deeper issue is incentives. Universities get the research culture they reward. If promotion systems continue to emphasise publication counts above intellectual contribution, AI will simply accelerate existing weaknesses. More papers will be produced, but not necessarily more insight. If institutions instead reward originality, methodological rigour, policy relevance, replication, and societal impact, researchers will adapt to those incentives as well.

From this perspective, the future of academic publishing is not primarily a technology story. It is an institutional story. AI is merely exposing questions that have existed for years: Who controls knowledge? Who verifies quality? Who benefits from the rewards of scholarship? And who should bear the costs?

For African universities, these questions arrive at an unusually opportune moment. The continent has long been disadvantaged by a publishing system whose rules were largely written elsewhere. As those rules come under pressure, there is an opportunity not merely to participate more fully in global scholarship but to help shape the next model.

Whether that opportunity is realised will depend less on artificial intelligence than on institutional imagination. The universities that thrive will not be those that use AI to produce the largest number of papers. They will be those that use it to build more trusted, more accessible, and more credible systems for creating and sharing knowledge.