The Hidden Price of Digital ID: Who Gets Shut Out?

When a government or a big corporation rolls out a shiny new digital ID system, the pitch is always the same: speed, security, inclusion. But on the ground, the story gets messier. I’ve spent years watching well-intentioned systems systematically lock out the people they were supposed to help. It’s not just a tech failure. It’s a failure of imagination, of empathy, and of design.

Digital identity systems aren’t neutral tools. They’re built on a stack of assumptions about who a person is, what they own, and how they live. For many of us, a smartphone, a fixed address, and a birth certificate are just facts of life. For millions of others, those things are either unaffordable luxuries or bureaucratic dead ends. The gap between the imagined user and the real human being is where exclusion quietly takes hold.

The Architecture of Exclusion

To see how these systems fail, you have to look at what they demand upfront. Most require some mix of the following: a government-issued foundational document, a mobile phone, a steady internet connection, biometrics that fit the algorithm’s expectations, and a permanent address. Each demand is a hurdle. Stack them up, and they become a wall.

Foundational Documents: The First Impossible Step

In many countries, getting a birth certificate or national ID card is a maze in itself. Rural communities, indigenous groups, and people displaced by war or climate disasters often have no papers at all. A digital ID system that asks for a physical ID as its starting point doesn’t fix the problem of invisibility—it just digitizes it. The World Bank puts the number of people without official identification at close to one billion. For them, a digital ID that’s gatekept by a paper one is a circle with no way in.

Picture a woman in a remote village who was never registered at birth because her family couldn’t afford the trip to a registration center. She’s lived her whole life without formal recognition. Then a humanitarian agency shows up with a tablet, promising digital IDs so she can get aid. The first field on the form? “National ID Number.” She’s already locked out before she can even begin.

The Smartphone Assumption

Digital ID schemes often lean on mobile apps or SMS codes. That assumes everyone has their own phone, reliable signal, and enough digital know-how. In low-income households, a single phone might be shared by an entire family. Women are especially likely to be on the losing end of that arrangement. A 2023 GSMA report found that women in low- and middle-income countries are 19% less likely than men to use mobile internet. When an ID system ties a person’s identity to a device, it ties their access to the power dynamics inside a home. A woman who shares a phone with her husband may not control her own authentication codes.

Even when a phone is present, biometric apps can choke on older or cheaper hardware. Software updates, cramped storage, and grainy cameras can make a person ineligible. The system logs a technical error. I see a person being told, silently, that they don’t count.

Biometric Barriers: When Bodies Don’t Fit the Mold

Fingerprint scanners, iris cameras, and facial matching algorithms are trained on statistical norms. Those norms usually come from datasets that tilt young, male, and able-bodied. Manual laborers can have fingerprints worn smooth. Older adults may have irises clouded by cataracts. People with certain disabilities might not be able to position their bodies the way the machine demands. A 2021 study in IEEE Transactions on Technology and Society found fingerprint failure rates as high as 15% for agricultural workers in India. These aren’t rare edge cases. They’re whole communities.

Facial recognition adds another layer of trouble. It misidentifies people with darker skin at higher rates, a bias that researchers like Joy Buolamwini have documented in painful detail. When a digital ID system uses facial matching as a gatekeeper, it amplifies the racial and gender biases already baked into the training data. A transgender person whose appearance doesn’t match the photo on file may get flagged as fraudulent again and again. The system doesn’t see a person in transition. It sees a mismatch to reject.

Connectivity and the Urban Bias

Many digital ID systems need real-time checks against central databases. That assumes always-on, high-speed internet. In rural stretches, refugee camps, and disaster zones, connectivity is patchy at best. Offline modes sometimes exist, but they’re often treated as afterthoughts, buggy and prone to sync errors. A seasonal worker who travels for a living might find their ID useless in a new region because the local system can’t reach the home database. Their identity becomes geographically conditional.

When Identity Becomes a Gatekeeper

Getting locked out of a digital ID system isn’t some abstract inconvenience. It’s a concrete loss. Without a recognized digital identity, a person may be unable to open a bank account, register a SIM card, vote, see a doctor, enroll in school, or collect social benefits. In countries that have moved public services online, the unregistered become non-citizens in practice, even if the law says otherwise.

During the COVID-19 pandemic, plenty of governments pushed emergency cash relief through digital channels. People without digital IDs got nothing. Some were also shut out of vaccine registration portals. The very tools meant to shield populations deepened the vulnerability of those already on the margins.

The Myth of the Universal User

Tech designers often talk about a “universal user”—a rational, able-bodied, literate person with a smartphone and a stable life. That person is a ghost. The real world is full of people in flux: refugees, homeless individuals, survivors of domestic violence, people with mental health conditions, and those who simply can’t afford the latest device. When you build a system for the universal user, you’re building for a fiction.

I’ve talked to aid workers who described digital ID systems that demand complex passwords from populations with low literacy or cognitive disabilities. I’ve read reports of systems that insist on a permanent address, but the users are nomadic pastoralists. In every case, the designers were genuinely surprised by the exclusion. They just hadn’t imagined these lives.

Administrative Violence

There’s a term I keep coming back to: administrative violence. It names the harm done by bureaucratic systems that ignore human complexity. When a digital ID system denies someone food, medicine, or shelter because of a data mismatch, that’s administrative violence. It’s not intentional cruelty, but it’s cruelty all the same. The system is tuned for efficiency, not kindness. And efficiency without kindness is a machine that grinds people down.

Data Protection and the Vulnerable

Even when vulnerable people get enrolled, they face outsized risks from data misuse. Digital ID systems hoover up sensitive personal information: biometrics, location trails, family links, financial histories. For someone fleeing domestic violence, a leak of their address data can be deadly. For a political dissident, a government ID database can morph into a tool of repression. Marginalized groups have more to lose from data breaches and less power to demand a fix.

Consent is another shaky idea. When access to essential services is conditioned on enrollment, consent isn’t freely given. A hungry person will press their thumb to any scanner. A refugee will share their iris pattern if it means shelter. The power gap between the enrolling authority and the individual strips consent of meaning. We need to stop pretending these are voluntary transactions.

Designing for the Margins

If we’re serious about inclusion, we have to flip the design process. Start with the most marginalized person, not the ideal one. Ask: Can someone with no documents enroll? Can someone without a phone authenticate? Can someone with unreadable fingerprints prove who they are? If the answer is no, the system isn’t inclusive—it’s exclusive with a marketing budget.

There are a few emerging models that offer some hope. Community-based digital identity systems, where trusted local organizations vouch for individuals, can sidestep the foundational document trap. Biometric systems that allow multiple modalities—voice, face, finger, iris—give people a fallback when one modality fails. Offline-first architectures, built for intermittent connectivity, respect the reality of infrastructure. But these approaches need real investment and political will. They’re not the default.

Legal Identity vs. Digital Identity

We also have to untangle legal identity from digital identity. Legal identity is a right recognized in international law: every person should be registered at birth and have access to official documentation. Digital identity is a tool. When we mush the two together, we risk making a person’s legal existence dependent on a tech platform. That platform might be owned by a private company, governed by terms of service, and driven by profit motives. A right should never hang on a user agreement.

What Accountability Looks Like

Governments and companies that deploy digital ID systems have to be held accountable for exclusion. That means publishing disaggregated data on who gets left out—by gender, age, disability, geography, and income. It means setting up independent redress mechanisms where people can challenge denials and fix errors without paying bribes or traveling for days. It means conducting human rights impact assessments before launch, not after the damage is done.

Civil society has a critical role here. Organizations that work directly with vulnerable communities can document exclusion, push for design changes, and offer parallel support. But they need funding and access. Too often, the conversation about digital ID is dominated by technologists and government officials. The voices of those affected show up as anecdotes, not as authorities.

Questions We Must Ask

Before any digital ID system goes live, I believe we have to ask a set of hard questions. Who won’t be able to enroll, and why? What happens to them? Is there a non-digital pathway to the same services? Who owns the data, and who can revoke access? What’s the remedy for a false rejection? If these questions can’t be answered with specificity and compassion, the system isn’t ready.

We also have to ask about power. Digital ID systems centralize control over identity. That control can be used to include, but it can also be used to surveil, exclude, and punish. The same database that delivers benefits can deliver arrest warrants. The design choices we make today will shape the political possibilities of tomorrow. We’re building infrastructure for a future we can’t fully predict, and we have to build it with humility.

FAQ

Why do digital ID systems require a pre-existing paper ID?

Many digital ID systems are built on top of existing national ID programs. They use the paper ID as a “breeder document” to verify a person’s identity before issuing a digital credential. This approach assumes that everyone already has a foundational document, which isn’t true for nearly a billion people globally. Without alternative enrollment pathways, these systems simply digitize existing exclusion rather than solving it.

How does biometric authentication fail for certain groups?

Biometric systems rely on patterns in fingerprints, irises, or faces. Manual laborers often have worn or damaged fingerprints that scanners can’t read. Older adults may have irises clouded by cataracts or skin that has lost elasticity, degrading fingerprint quality. Facial recognition systems have documented higher error rates for people with darker skin tones, women, and transgender individuals whose appearance may not match outdated photos. When a system offers no alternative to biometrics, these groups are effectively locked out.

What can be done to make digital ID systems more inclusive?

Inclusion requires designing for the most marginalized from the start. This means offering multiple enrollment pathways, including community-based vouching for those without foundational documents. It means supporting offline authentication and multiple biometric modalities with fallback options. It also requires strong data protection laws, independent oversight, and accessible grievance mechanisms. Most importantly, non-digital alternatives must remain available so that no one is forced to use a system that doesn’t work for them.

Is digital identity a human right?

Legal identity—being recognized as a person before the law—is a human right, as affirmed by the UN Sustainable Development Goals. Digital identity is a tool that can support legal identity, but it is not a right in itself. When digital identity becomes mandatory for accessing basic services, it effectively becomes a de facto requirement for exercising rights. This blurring of tools and rights is dangerous and must be resisted through policy and advocacy.

The conversation about digital ID is often framed in terms of progress. But progress that leaves people behind isn’t progress—it’s a redesign of inequality. As we build these systems, we have to remember that identity is not a data point. It’s a human story, complex and sacred. And no story should be erased by a login screen.