The Quiet Violence of Digital ID: How Systems Fail the Most Vulnerable

Talk to a government minister or a tech vendor about digital identity, and you’ll hear a lot about efficiency, modernization, and frictionless access to services. The pitch is seductive: biometric databases and mobile authentication can cut through bureaucracy, reduce fraud, and finally bring the unbanked into the formal economy. But if you spend time with the people these systems are supposed to rescue—refugees, the elderly, those without fixed addresses, survivors of domestic violence, people with disabilities—a very different picture emerges. For many of them, digital ID isn’t a bridge to inclusion. It’s a locked door, and sometimes a weapon.

I’ve spent years examining how identity verification systems intersect with social inequality. The pattern is stubbornly consistent across continents. The same populations that are promised protection end up being the ones most reliably excluded. The reasons aren’t glitches or bad luck. They’re structural, baked into the design, and they deserve a far more honest accounting than the glossy brochures ever provide.

The Myth of Universal Access

Digital identity is usually sold as a universal good. The World Bank’s ID4D initiative tells us that nearly a billion people lack official identification, and the proposed fix is digital: biometric databases, mobile authentication, centralized registries that can verify a person’s existence with a fingerprint or an iris scan. The logic has a clean, seductive symmetry. No paper birth certificate? A digital record fills the gap. Can’t travel to a government office? A mobile enrollment unit comes to you.

But this logic depends on a world that doesn’t exist. It imagines that every person has a stable, legible identity that can be neatly captured and stored. It assumes electricity, internet connectivity, and compatible devices are spread evenly across the population. It assumes the institutions running these systems are trustworthy, and that the people being enrolled have no reason to fear them. None of these assumptions hold for the populations most in need of protection. And when the assumptions collapse, the system doesn’t just fail—it punishes.

Who Gets Left Behind

The exclusion isn’t random. It follows the fault lines that already cut through society. People who are poor, displaced, or living with disabilities are the most likely to be locked out—and then punished for their exclusion.

People Without Fixed Addresses

Many digital ID systems demand proof of residence. For someone living in a slum, a temporary shelter, or on the street, that requirement is a wall. India’s Aadhaar system—the largest biometric ID program on the planet—gets praised for its scale and ambition. But researchers have documented, again and again, how homeless individuals are turned away at enrollment because they can’t produce an address. Without an Aadhaar number, they’re cut off from subsidized food, cooking gas, and other welfare benefits. The ID that was supposed to include them becomes the precise instrument of their exclusion.

Survivors of Domestic Violence

Digital ID systems often link family members together in ways that are not just inconvenient but dangerous. In some implementations, a woman’s identity is tethered to her husband’s, meaning she can’t access benefits or services without his authentication. For a woman fleeing an abusive partner, that linkage can be life-threatening. Changing a registered address or delinking from a spouse may require in-person visits, official documentation of separation, or even the husband’s consent. These aren’t realistic options for someone in hiding. The system, designed for a stable nuclear family, becomes another chain.

People with Disabilities

Biometric authentication is often pitched as the most inclusive form of identification because it doesn’t require literacy. That pitch ignores the many ways biometrics can fail. Fingerprint scanners don’t work for people with worn fingerprints—common among manual laborers and the elderly. Iris scanners can be inaccessible for people with visual impairments or certain eye conditions. Facial recognition systems perform poorly on people with facial differences or those who can’t hold the required pose. When the system can’t capture a “readable” biometric, the person is often left unenrolled, with no meaningful alternative pathway. The promise of inclusion evaporates at the scanner.

Stateless and Displaced Populations

Refugees and stateless people are perhaps the most systematically excluded. They often lack the foundational documents—birth certificates, national ID cards—that digital systems demand as proof of eligibility. Humanitarian agencies have piloted digital ID programs in refugee camps, but these IDs are rarely recognized by host governments. A refugee may have a UNHCR-issued digital identity, but if the local bank or mobile network operator won’t accept it, the ID is functionally useless. The result is a two-tier system: a durable digital identity for citizens, and a fragile, limited one for the displaced. The people who need recognition the most get the version that opens the fewest doors.

The Hidden Costs of Biometric Failure

When a biometric system fails to recognize a person, the consequences go far beyond inconvenience. In many countries, digital ID is now the gateway to essential services: food rations, healthcare, pensions, even the right to vote. A failed fingerprint scan at a ration shop can mean no food for a family that week. A mismatched iris scan can block a pension payment for months. These aren’t hypothetical scenarios. They’re documented, recurring realities in countries that have rushed to digitize their welfare systems.

Take India’s Public Distribution System, which distributes subsidized grain to millions of poor households. After Aadhaar-based biometric authentication was mandated at ration shops, researchers found that technical failures—scanner malfunctions, connectivity issues, biometric mismatches—led to significant numbers of legitimate beneficiaries being turned away. The elderly and manual laborers were hit hardest because their fingerprints were often unreadable. Some were told to come back another day. But for those who had traveled long distances or couldn’t afford to miss a day’s wages, “come back later” wasn’t a viable option. It was a quiet denial of food.

These failures aren’t glitches. They’re predictable outcomes of systems designed without accounting for the full range of human bodies and human circumstances. When a system is built around an idealized user—someone with stable housing, readable fingerprints, a smartphone, and the physical ability to travel to enrollment centers—it will inevitably fail those who deviate from that norm. And the people who deviate are, overwhelmingly, the people who most need the services the ID is meant to unlock.

When Exclusion Becomes Structural Violence

There’s a term for this: structural violence. It describes the way social structures and institutions harm people by preventing them from meeting their basic needs. When a digital ID system systematically excludes certain groups from food, healthcare, or legal recognition, it’s not a technical error. It’s a form of violence embedded in the design.

This violence is often invisible to the people who design and implement these systems. Engineers and policymakers tend to see exclusion as an edge case—a problem to be solved with better technology, more enrollment drives, or exceptions processes. But the edge cases aren’t marginal. They’re central to the question of whether a system is just. A system that works for 95% of the population but starves the other 5% isn’t a 95% success. It’s a machine for producing suffering, and its designers bear responsibility for that suffering.

The language we use matters here. When a person is denied food because their fingerprint doesn’t scan, we shouldn’t say they “failed authentication.” We should say the system failed them. The burden of failure must rest on the system, not on the human being whose body doesn’t conform to the system’s narrow expectations.

Privacy as a Precondition for Safety

Exclusion isn’t the only risk. For many vulnerable populations, the very act of being registered in a digital ID system creates new dangers. Centralized identity databases can be used to track, profile, and target individuals. For refugees, having their biometrics stored in a database controlled by a host government—or shared with their country of origin—can put them and their families at risk. For survivors of domestic violence, a digital trail of addresses and family linkages can be exploited by abusers. For political dissidents, a national ID system tied to biometrics can become a tool of surveillance and repression.

Privacy isn’t a luxury for the vulnerable. It’s a precondition for safety. Yet digital ID systems are often designed with the assumption that all users want to be visible to the state, that visibility is inherently beneficial. For many people, visibility is dangerous. A truly inclusive system would allow for anonymity and pseudonymity where needed, and would give individuals control over how their identity data is shared and with whom. Few existing systems offer anything close to that.

What a Principled Approach Looks Like

If we’re serious about building identity systems that don’t harm the most vulnerable, we need to start from a different set of assumptions. Here are the principles that should guide any digital ID initiative claiming to serve the public good.

1. Voluntary Enrollment with Meaningful Alternatives

No one should be coerced into enrolling in a digital ID system by making it the only way to access food, shelter, or medical care. There must always be an alternative, analog pathway to essential services. The alternative must be equally accessible, not a bureaucratic labyrinth designed to push people toward the digital option.

2. Designed for the Margins, Not the Mainstream

Systems should be built to accommodate the hardest-to-reach populations first. If a system works for a homeless person with worn fingerprints and no smartphone, it will work for everyone. This means offering multiple authentication methods, accepting a wide range of foundational documents, and providing in-person support that meets people where they are.

3. Data Minimization and User Control

Collect only what is necessary. Store only what is essential. Give individuals the ability to view, correct, and delete their own data. Allow people to use pseudonyms or limited-purpose credentials when full identification isn’t required. Consent must be informed, specific, and revocable.

4. Independent Oversight and Redress

No identity system should operate without independent, well-resourced oversight. There must be clear, accessible mechanisms for people to challenge errors, appeal exclusions, and seek compensation for harms caused by system failures. These mechanisms must be designed with the needs of the most vulnerable users in mind—not as an afterthought.

5. Context-Specific Design

There is no one-size-fits-all digital ID. A system that works in Estonia won’t work in a refugee settlement in Kenya. Design must be grounded in the specific social, political, and infrastructural realities of the population it serves. This requires genuine, sustained engagement with communities—not just consultations with elites and tech vendors.

The Cost of Getting It Wrong

When digital ID systems exclude, the consequences cascade. A person without recognized identity can’t open a bank account, register a SIM card, vote, inherit property, or enroll their children in school. In countries where welfare systems are tied to digital ID, exclusion means hunger, untreated illness, and homelessness. These aren’t abstract policy failures. They’re material, bodily harms.

We also need to talk about the international dimension. Donor agencies and development banks are pouring billions into digital ID projects across the Global South, often with the stated goal of “leaving no one behind.” But the procurement contracts go to many of the same technology vendors who have built exclusionary systems elsewhere. The incentives are misaligned: vendors profit from building and maintaining systems, not from ensuring those systems serve the most marginalized. Governments profit from the ability to track and control populations. The people who bear the costs of failure have the least power in the design process.

This isn’t a call to abandon digital identity. Well-designed systems can reduce bureaucratic friction, lower corruption, and give people portable proof of who they are. But the gap between what’s possible and what’s being built is wide and growing. Closing that gap requires a fundamental shift in who gets to define the problem and who gets to design the solution.

Frequently Asked Questions

Why are digital ID systems considered important for development?

Proponents argue that digital ID systems can help people without formal identification access banking, healthcare, education, and government services. They can also reduce fraud and improve the efficiency of public administration. However, these benefits depend on systems being designed inclusively and implemented with strong safeguards—conditions that are often not met in practice.

What makes biometric identification particularly risky for vulnerable groups?

Biometric systems rely on the assumption that every person has stable, readable biological markers. In reality, manual laborers may have worn fingerprints, elderly people may have faded irises, and people with certain disabilities may not be able to provide the required biometric sample. When the system fails to capture or match biometrics, these individuals are often excluded from services, with no easy recourse.

Can digital ID systems be designed to protect privacy?

Yes, but it requires deliberate architectural choices. Decentralized models, where individuals control their own identity data and share it selectively, offer more privacy protection than centralized databases. Systems can also be designed to support anonymous or pseudonymous credentials for services that do not require full identification. However, these approaches are rarely adopted because they conflict with state interests in surveillance and control.

What should I look for to determine if a digital ID system is rights-respecting?

Key indicators include: whether enrollment is voluntary, whether non-digital alternatives exist for essential services, whether the system collects only minimal data, whether users can access and correct their own data, whether independent oversight bodies have real power, and whether marginalized communities were meaningfully involved in the design process. If the answer to any of these is no, the system likely poses significant risks to vulnerable populations.