The Hidden Cost of Digital ID: Who Gets Left Out

Governments and corporations are racing to build digital identity systems, selling them as the fix for fraud, inefficiency, and exclusion. The pitch is tempting: one verifiable digital credential that opens the door to banking, healthcare, voting, and social services. For many, it really is a convenience. But for a huge, often invisible slice of the population, digital ID isn’t a bridge to inclusion. It’s a new wall, reinforced by technology.

My work has taken me into the messy, human corners these systems ignore. I’ve sat with elderly people baffled by biometric scanners, talked to homeless individuals who’ve lost their documents for the dozenth time, and listened to domestic violence survivors who fear that a unified digital trail will lead an abuser straight to their new home. The conversation around digital identity is obsessed with speed and security. We need to shift it toward a different value: equity. Without that, we’re not building a better system. We’re just digitizing old exclusions and inventing fresh ones.

The Myth of the Universal User

Every digital ID system rests on assumptions about the people who will use it. The most dangerous is the myth of the universal user: someone with a fixed address, a smartphone, reliable internet, a formal education, and a body that fits standard biometric templates. This person doesn’t represent a huge portion of humanity. When we design for this fictional archetype, we hard-code exclusion into the system’s foundation.

Consider the first step: enrollment. Many digital ID programs demand an existing foundational ID, like a birth certificate, just to get started. That creates a circular trap for the estimated one billion people worldwide who lack any official identity document. If you were born in a remote village without a formal hospital, or you’re a refugee whose papers were destroyed, or you belong to a marginalized ethnic group the state has systematically denied citizenship, you can’t even begin the process. The digital system doesn’t just fail to include you; it actively re-marginalizes you by making your lack of official existence a permanent, tech-enforced condition.

The Biometric Bottleneck

Biometrics are often sold as the ultimate inclusive identifier because, the argument goes, everyone has a body. That’s a dangerously simplistic view. In practice, biometric systems fail predictably for certain groups. Manual laborers have worn-down fingerprints that scanners can’t read. People with arthritis, Parkinson’s, or cerebral palsy may not be able to hold their hands or heads still enough for a successful capture. Cataracts and other eye conditions defeat iris scans. A system that demands a perfect biometric match doesn’t have a minor glitch; it systematically excludes the elderly, the disabled, and the working poor.

I once reviewed a pilot program that used fingerprint scanners to disburse pension payments in a rural area. It worked smoothly for most, but it consistently failed for a group of elderly women whose fingers had been worn smooth by decades of hand-rolling incense sticks. The digital system, in its rigid perfection, declared them unidentifiable. Their pensions were suspended. The very tool meant to ensure they received their due rendered them invisible. This isn’t a bug. It’s a design failure that confuses a fingerprint with a person.

The Connectivity and Literacy Gap

Even when someone has the right documents and readable biometrics, the digital divide can still lock them out. Many digital ID systems depend on smartphone apps, online portals, or real-time verification. That assumes steady electricity, affordable data, and reliable network coverage. For rural communities, the urban poor who share a device intermittently, or anyone living where infrastructure is fragile, this assumption is a fantasy. A digital ID that needs a smartphone and a strong signal isn’t an ID for everyone. It’s an ID for the connected class.

Then there’s digital literacy. Navigating a complex online portal, understanding authentication steps, or simply typing a password can be a steep barrier for older adults, people with limited schooling, or those with cognitive disabilities. When a digital ID system replaces a staffed service counter with a website, it doesn’t just save money. It cuts off access for anyone who can’t manage the digital interface. The result is a two-tier reality: a fast, convenient digital lane for the tech-savvy, and a slow, degrading, or nonexistent analog lane for everyone else.

The Gendered and Social Risks of Digital Identity

Digital ID design rarely accounts for the specific dangers faced by women and marginalized groups. A unified digital identity that links a person’s data across health, banking, social welfare, and location can become a tool for control and surveillance. For a woman fleeing domestic violence, a digital trail that exposes her new address, her financial transactions, or her children’s school enrollment is a serious threat. The same seamlessness that makes digital ID efficient for the state can make it deadly for the vulnerable.

Data security and consent are equally troubling. In many digital ID programs, the individual has little to no say over who accesses their data and why. Someone applying for a subsidized food ration may not realize their identity data could be shared with law enforcement or immigration authorities. For undocumented migrants, this creates an impossible choice: claim a basic right to food and risk deportation, or stay invisible and go hungry. A system built without a strong, user-controlled consent framework isn’t an empowerment tool. It’s a surveillance machine that feeds on the desperation of the marginalized.

The Fallacy of One-Size-Fits-All

The push for a single, foundational digital ID often comes from a desire for administrative tidiness. But human identity isn’t tidy. It’s contextual, relational, and fluid. We show different sides of ourselves to a doctor, a banker, a friend, and a voter. A system that collapses all these contexts into one universal identifier creates a single point of failure. If that ID is compromised, stolen, or revoked, the person doesn’t just lose access to one service. They are erased from the formal system entirely. For someone already living on the edge, this is a catastrophic risk.

Worse, the drive for digital ID often comes with a push to make it mandatory for essential services. That turns a tool into a gatekeeper. When a digital ID is required to buy a SIM card, open a bank account, or receive a welfare payment, those who can’t obtain or maintain the ID aren’t just inconvenienced. They are actively shut out of economic and social life. The system punishes the very people it claims to serve.

Designing for the Margins: A Principled Path Forward

So, what does a just digital identity system look like? It starts by ditching the myth of the universal user and designing for the margins instead. That means asking, “Who is most likely to be excluded by this technology, and how do we build the system around their needs first?” It requires a shift from a purely technical mindset to a socio-technical one, where the human context isn’t an afterthought but the primary design constraint.

1. Multi-Modal and Fault-Tolerant Systems

A just system can’t rely on a single mode of authentication. Biometrics should be one option among many, alongside passwords, PINs, physical tokens, and—most importantly—a well-funded, dignified process for exception handling. When a fingerprint scanner fails, the fallback shouldn’t be a bureaucratic nightmare that demands a 200-mile trip to a central office. It should be a local, human-mediated process that respects the person’s time and dignity. The system must be designed to fail gracefully, with the human as the ultimate authority on their own identity.

2. Offline Functionality and Analog Equivalents

Digital systems must have offline capabilities. A smart card with a chip that can be verified with a simple, battery-powered reader is far more inclusive than a cloud-dependent biometric system. In addition, no essential service should be accessible only through a digital ID. A parallel, non-digital pathway must always exist, and it must be equally efficient and respectful. Otherwise, the analog option becomes a second-class queue for the poor and the elderly, and the digital ID becomes a tool of coercion.

3. User-Controlled, Contextual Data Sharing

Identity systems should be built on principles of data minimization and user consent. A person should be able to prove they are over 18 without revealing their exact birth date, or prove they are eligible for a health service without revealing their full medical history. Technologies like zero-knowledge proofs offer a technical path to this, but the real requirement is a legal and governance framework that puts the user in control. The default setting for any data sharing should be “no,” and the process for giving consent must be genuinely understandable, not buried in pages of legalese.

4. Inclusive Governance and Redress

Finally, the people most affected by a digital ID system must have a seat at the table when it is designed, implemented, and audited. This means going beyond token consultations and ensuring that civil society organizations, disability rights advocates, groups representing the elderly, and organizations working with the homeless have real power in the governance structure. There must also be a free, accessible, and rapid mechanism for individuals to challenge errors, exclusion, or misuse of their data. A system without accountability is a system designed to abuse power.

Frequently Asked Questions

Why can’t people without an ID just get one?

The barriers to obtaining a foundational ID are often deeply systemic. They can include lack of birth registration in remote areas, the cost and distance of traveling to registration centers, discriminatory laws that deny citizenship to certain ethnic or religious groups, and the loss of documents due to conflict or disaster. A digital ID system that requires an existing foundational ID simply digitizes these pre-existing exclusions rather than solving them.

Are biometrics really that unreliable for some people?

Yes. The failure rates for biometric systems are not uniform across the population. Manual laborers, the elderly, people with certain disabilities, and even people with skin conditions can experience much higher failure-to-capture or failure-to-match rates. A system that treats a 1% failure rate as acceptable engineering is still failing millions of people, and those people are disproportionately from already marginalized groups.

What is the alternative to a single, universal digital ID?

The alternative is a decentralized, user-centric identity model. Instead of one monolithic ID number, individuals could have multiple, context-specific credentials that they control. For example, a person could use a verifiable credential from a health clinic to access medical services without revealing their full identity, and a separate credential from a utility company to prove their address. This approach limits the damage from a data breach, reduces surveillance risks, and allows people to manage their identity in a way that fits their complex lives.

How can digital ID systems be made safer for survivors of domestic violence?

Safety for survivors requires that digital ID systems are not designed as a single, unified record of a person’s life. There must be mechanisms to decouple or mask data trails—for example, allowing a survivor to use a service-specific credential that does not reveal a new address. Systems must also have strict access controls and audit logs, so that any unauthorized access by an abuser can be detected and stopped. Most importantly, the design process must include input from domestic violence experts and survivors to identify risks that technologists would never see on their own.

The question isn’t whether we should have digital identity systems. The question is what values we embed in their code and their governance. A system optimized for efficiency and control will inevitably exclude and endanger the most vulnerable. A system optimized for equity and human dignity can be a genuine tool for inclusion. The choice is ours, and it is a political one, not a technical one. We must demand that our digital infrastructure be built not for the mythical average user, but for the real, complex, and precious diversity of human life.