The People Institutions Cannot See: Why Records, Demolitions, and Reparations Belong to the Same Moral Problem

What does a medical record have in common with a demolished school in Makoko or a war survivor waiting decades for recognition?

At first, almost nothing. One belongs to healthcare technology, another to urban governance, and the third to postwar justice. Yet all three turn on the same hidden question: What happens to people whose realities do not fit the system’s record of them?

A patient’s crucial history may be buried in unstructured notes. A waterfront family may have lived for decades in a community that the law refuses to recognize as a legitimate settlement. An amputee may carry an undeniable physical record of war, yet remain absent from the budget, the healthcare system, and the nation’s public memory.

The common problem is not simply inefficiency or bureaucratic delay. It is a failure of institutional visibility. When a system cannot retrieve, classify, or acknowledge a person’s reality, that person becomes easier to exclude. The result is a peculiar form of vulnerability: people are present in the world but missing from the machinery designed to respond to them.

The deepest lesson is this: administrative records are not neutral descriptions of reality. They are infrastructure for distributing attention, protection, and justice.

The First Injustice Is Often Invisibility

Consider the medical record. Healthcare organizations increasingly need to retrieve information from documents that were not designed for easy use: handwritten notes, scanned reports, fragmented histories, forms, correspondence, and other unstructured records. The promise of claim automation and automated medical record review is usually expressed in the language of time and cost. Find the relevant data faster. Reduce manual review. Improve consistency. Process more claims with fewer resources.

Those are practical benefits, but they conceal a moral dimension. A record review system determines which facts become actionable. It can identify a diagnosis, a prior injury, a treatment, or a contradiction that would otherwise remain buried. In doing so, it does not merely accelerate paperwork. It changes the probability that a person will be believed, treated, compensated, or denied.

The same logic appears in urban clearance. Lagos authorities described demolitions in Makoko through categories such as safety zones, power lines, and legal land rights. Residents described something else: homes, schools, canoe routes, fish smoking, trading networks, and a place they had known as home for generations. Both descriptions may contain facts, but they do not contain the same facts. One renders the area as a hazard and an unlawful occupation. The other renders it as a functioning social and economic system.

What the official map leaves out is not incidental. It is precisely what makes displacement devastating.

A house is not only a structure. It may be the location from which a person reaches a market, stores equipment, sends a child to school, receives medical care, and maintains ties with relatives. Destroy the structure and the visible loss is one building. Destroy the network attached to it and the actual loss becomes education, income, health, safety, and social belonging.

The woman who returned from the market to find that her home had vanished did not merely lose property. She lost the continuity that allows a life to remain intelligible from one day to the next. Her family was suddenly forced into survival mode, sleeping in the rain, washing clothes at night and wearing them wet, and trying to keep a sick child alive.

A system that records only the structure can report clearance. A system that records the life around the structure must report displacement.

The first question in any administrative process should not be, “What category does this person occupy?” It should be, “What reality will disappear if our category is incomplete?”

From Recognition to Repair

Sierra Leone’s experience after its civil war reveals the same problem across a longer time horizon. The conflict ended in 2002, but many amputees and war wounded continued living in severe poverty, sometimes in deteriorating homes built by foreign donors and sometimes reduced to begging. Their bodies preserved the most visible signs of the war, yet visibility alone did not guarantee support.

This is an important distinction. A person can be publicly visible and institutionally invisible at the same time.

The amputees were not hidden in the literal sense. Their injuries were unmistakable. But they were not adequately present in the systems that provide housing, healthcare, education, social security, and reparations. The country could remember the war in broad historical terms while failing to maintain the everyday records required to care for its survivors.

That is why an apology and a national Remembrance Day matter, but also why they are insufficient. Symbolic recognition answers the question, “Do we acknowledge what happened?” Structural repair answers a harder question: “Will that acknowledgment change what happens next?”

A government can declare a holiday, issue an apology, and announce grants. Yet if eligible survivors cannot access healthcare, if children cannot receive educational support, or if people excluded by earlier programs have no clear appeal process, recognition remains ceremonial. The institution has named the wound without building a mechanism for treating it.

This is where the language of medical records becomes unexpectedly useful. A claim is not resolved because relevant evidence exists somewhere. It is resolved when the evidence is retrievable, connected to the correct person, interpreted in context, and linked to a decision process.

The same four conditions define meaningful reparations:

1. Retrievability: Can the state identify survivors and their needs?
2. Context: Does it understand how an injury affects income, family responsibilities, health, and education?
3. Decision linkage: Are those facts connected to grants, care, housing, or social security?
4. Appealability: Can a person challenge an omission or denial?

Without these conditions, a promise is like a medical fact trapped in an unreadable file. It may exist, but it cannot reliably help anyone.

The public commitments to prioritize war wounded people and their children, provide free healthcare, and correct past exclusions are therefore significant only if they become durable administrative pathways. A survivor should not have to repeatedly prove the same injury to different offices. A child should not lose an educational opportunity because a parent’s status was never properly recorded. A person denied care should know where to appeal and receive an answer.

Justice is not complete when the state expresses regret. It is complete when regret becomes an operating system.

The Danger of One Sided Legibility

Institutions need categories. Electricity providers need safety corridors. Cities need land rules. Insurers and hospitals need claims procedures. No society can function without records, boundaries, and standards.

The problem begins when only the institution’s categories count as real.

In Makoko, residents reportedly understood an initial agreement as a clearance of about 30 meters around the power lines. They later saw more than 250 meters cleared. That gap is not merely a disagreement over measurement. It is a collapse of procedural trust. The affected people could not tell which rule governed their lives, when it would be enforced, or whether the stated safety rationale was being applied consistently.

When the operating rule changes without written notice, consultation, or a visible explanation, a legal process can feel indistinguishable from arbitrary force. The issue is not whether safety concerns are genuine. Structures beneath high tension power lines may pose real danger. The issue is whether safety is being used as a narrow protection for human life or as a broad administrative instrument that also clears valuable land and weakens a community’s bargaining position.

Uncertainty about private real estate interests intensifies this suspicion. The claim may not be authenticated, but the suspicion itself has a cause: repeated demolitions, unclear communication, and the absence of a transparent account of what will happen to the land afterward. When public decisions produce private advantages and residents receive little explanation, people reasonably question whose interests the official record serves.

This suggests a useful framework: legibility must be reciprocal.

The state wants residents to be legible. It wants names, addresses, permits, maps, and proof of eligibility. But residents also need the state to be legible. They need to know the rule, the boundary, the timeline, the decision maker, the evidence, the relocation plan, and the remedy if officials fail to keep their promises.

A one sided system says, “We can identify you whenever we need to regulate you.” A reciprocal system says, “You can also identify the rules and obligations governing us.”

That principle applies to automated medical review as well. Automation can reduce the burden of finding information, but it can also make decisions harder to question if the process becomes opaque. A patient or claimant should not be trapped inside an invisible classification. Efficient systems need traceable evidence, clear explanations, human review for contested cases, and a way to correct incomplete records.

Speed without contestability is not progress. It is simply faster finality.

The Cost of Treating Social Life as Unstructured Data

The phrase “unstructured medical records” sounds technical, but it points to a broader social reality. Human lives are usually unstructured from the perspective of institutions. Families change addresses. Work is informal. Care is unpaid. Injuries affect multiple generations. Communities develop their own economies outside official plans. Trauma persists in bodies, habits, and lost opportunities long after an event has left the news.

Institutions prefer clean fields: age, address, diagnosis, entitlement, ownership, risk level. But the most consequential facts often live between those fields.

In Makoko, the destruction of homes also damaged schools and interrupted children’s education. It disrupted fishing, fish smoking, canoe transport, and small scale trade. Families reportedly survived on little more than gari and water. These consequences are not secondary effects. They are the main event viewed from the perspective of the people affected.

Likewise, an amputee’s need is not captured by the word “injured.” The injury may mean chronic pain, inability to perform certain work, dependence on relatives, higher household costs, stigma, and reduced educational prospects for children. If a reparations program records only the original injury but not its continuing social consequences, it will under repair the harm.

This is why automated retrieval, at its best, should not be understood as replacing human judgment with machine judgment. Its deeper purpose is to make complex evidence available for better judgment. The technology is valuable when it helps decision makers see patterns that fragmented records conceal, not when it gives them an excuse to avoid listening.

A useful model is the chain of institutional care:

Presence, evidence, interpretation, action, feedback.

A person must first be present in the system. Their relevant evidence must then be captured and retrievable. That evidence must be interpreted in context. Interpretation must lead to action. Finally, the person must be able to provide feedback and correct the record.

Break any link and care fails.

A Makoko resident may be physically present but absent from the city’s recognized housing plan. A war survivor may be known symbolically but absent from a functioning benefits register. A patient may have extensive documentation that no reviewer can efficiently locate. In each case, the issue is not a total absence of information. It is a failure to turn information into accountable response.

Designing Institutions That Can Remember

The practical response is not to reject rules, technology, or modernization. It is to design them around the people who bear their consequences.

For displacement, that means conducting a social impact inventory before demolition, not afterward. Authorities should document households, schools, businesses, health needs, transport routes, property claims, and community institutions. Notice should be written, specific, and delivered through channels residents actually use. The affected population should know the exact area covered, the reason for the boundary, the timetable, the relocation options, and the compensation process.

For reparations, it means creating a permanent survivor registry linked to healthcare, education, housing, grants, and social security. The registry should not become a surveillance tool or a bureaucratic obstacle. It should reduce repeated proof, protect privacy, and include an independent appeals mechanism. Reparations should be measured not by announcements but by outcomes: children enrolled in school, survivors receiving care, households with stable income, and neglected cases corrected.

For medical record automation, it means treating retrieval as the beginning of accountability rather than the end. Systems should show the evidence behind a result, flag uncertainty, preserve source documents, identify missing information, and route high consequence decisions to trained human reviewers. Efficiency should expand the time available for judgment, explanation, and compassion.

Across all three settings, one rule stands out: do not automate or enforce what you have failed to understand.

The best institution is not the one that produces the cleanest dashboard. It is the one that can detect when the dashboard is lying by omission.

Key Takeaways

1. Audit omissions, not only errors. Ask which people, consequences, and forms of evidence are absent from the official record.
2. Make rules reciprocal and visible. Provide written boundaries, timelines, reasons, responsible officials, and appeal channels before taking action.
3. Connect recognition to services. An apology, classification, or eligibility decision has value only when it reliably unlocks healthcare, education, housing, income support, or repair.
4. Preserve context in automation. Use technology to retrieve complex evidence, but require traceability, uncertainty flags, human review, and correction mechanisms.
5. Measure outcomes at the level of lived experience. Count restored schooling, accessible healthcare, stable livelihoods, and repaired homes, not merely processed cases or completed demolitions.

The central challenge is not that modern institutions lack information. It is that they often possess information in forms that protect the institution better than they protect the person.

A database can know that a building exists without knowing that a child’s school is inside it. A government can know that a war occurred without knowing how a missing limb reshapes a family’s future. A medical archive can contain the decisive fact without making it findable when a claim is reviewed.

To remember people institutionally means more than storing their names. It means preserving the connections between facts and consequences. It means ensuring that a person does not disappear simply because their life is too complex for the available form.

The future of humane governance will depend on this distinction. The question will not be whether institutions have become more digital, more efficient, or more orderly. It will be whether their systems can see the whole human cost of the decisions they make.

Because the people most at risk are often not those with no record at all. They are the people whose record contains only the part of their reality that power finds convenient.