Identity verification is, by design, a barrier. We use it to make sure that the right person gains access to information, services or money. But we also risk excluding the people who need the service the most.

The most important public services often have high numbers of users that are already excluded by society, or who have difficulty using digital services.

In this post, we focus on the challenges and practical steps governments can take to make digital identity inclusive.

If you can avoid the need to verify someone’s identity as a person, you should.

Maybe you just need a secure account, or perhaps just a one-time code which can be sent to users. Digital identity services are often treated as a magic bullet to solve all problems around sharing sensitive information, checking eligibility and giving out money. This is risky. You may be giving stakeholders confidence that you have “solved” those security challenges, but it is a false sense of confidence that hides insecure operational work arounds and the needs of excluded communities.

Attributes of identity, such as a standard set of “name, address, date of birth, gender,” do not conveniently map onto how all individuals live their lives. And these attributes can all change over time. When these are codified into digital identity services and treated as static, it creates barriers when something changes.

In many cultures people change their name when getting married. In another example of fluidity, in Native American traditions, names can change over a person’s life, from childhood to adolescence to adulthood, but also in recognition of changing life experiences and achievements”.

People also change gender, which often also causes a change in name. Some countries are moving away from using gender in identity systems altogether. In July 2020, in a letter written by the Education Minister to parliament, the Dutch government announced their plan to remove gender from identity cards in five years. A proposal by the government of New Zealand to the ICAO in 2012 suggested the removal of gender from international travel documents.

Even seemingly solid facts like date of birth can be uncertain or be revised in certain communities. For example, often temporary dates of birth are assigned to systems of record where proof is not available at the time, which can migrate onto official documentation, causing issues when the correct date of birth becomes available.

So it is important to check whether you really need to use every attribute as part of your identity system. If you can eliminate them, you eliminate complexity and barriers for users.

Short of eliminating attributes, you can ensure that not only are there legal rights to make changes, but the path for users to make those changes in their official identity is smoothed also. In Argentina, a law passed in 2012 gives citizens the right to have their chosen gender identity recognised explicitly, including the right to request new identity documents if their current documents do not have their recognised gender on them.

By doing the hard work to really understand the users and services that will need digital identity, you can reduce the likelihood identity verification becomes a means of exclusion. But it requires significant effort.

In-depth user research with marginalised groups is essential in identifying where identity could become a barrier rather than a gateway. A great example is New Zealand where the Māori community consider identity a treasured possession (Taonga), which as part of the Treaty of Waitangi they therefore own. Building trust with the Māori community that their identity will be managed appropriately is essential to ensuring they are able to access services equally with the rest of society. To ensure this happens there has been extensive research done with the Māori community on their reaction to digital identity. Furthermore, engaging and enabling Māori leaders to be part of the process to design their new digital identity service is explicitly called out in the Identity Trust Framework that is currently being developed by the New Zealand government.

This World Bank study on gender related barriers to identity is a great example of how working directly with a community can provide practical implementation advice to increase inclusion. As we have previously suggested, such engagement could ideally be included in the engagement around the creation of a trust framework.

Service teams should aim for equal outcomes, rather than a singular service. Trying to support all users to use the most sophisticated digital channels will result in exclusion.

You’ll need an answer for every edge case and all of the barriers, so the most important thing you can do to be inclusive is design and plan alternatives. This doesn’t mean you have to create completely parallel systems, or leave some users on a legacy system, but you do need to design a service that can handle a range of options for verifying identity.

For example, Universal Credit, the UK Government’s reform of its welfare system, is a fully digital service that must support a large number of users who normally suffer barriers in service provision due to digital and social exclusion. Given it serves over 5.8 million users, the service has to work hard to ensure it does not create barriers to access, given the diversity of circumstances of its users. For example, while most users access the service via the web, the service ensures that other means are offered to claim the benefit - JobCentre, telephone or home visits are available for those who cannot access the service online. These channels are all part of the same digital service.

This approach is also taken to identity verification for Universal Credit. An essential part of applying for the benefit involves verification of the user’s identity, but the reality of such a broad population means that it’s not realistic to expect everyone to do so through an online channel, and the service has evolved to offer a wide variety of means to users. Guidance published by the UK parliament shows the variety of identity verification that the service offers:

• online ID verification from multiple sources.

• face-to-face ID interviews where the user brings documentary evidence.

• a Know Your Customer-type process of gathering information that the user has and comparing with what the service trusts, for those who do not have traditional evidence.

• a process whereby service staff directly verify the user’s identity, if they are known by them personally, through the provision of other services at that location.

The important lesson here is that this variety of means has been provided within the same service offering.

India’s Aadhar biometric ID system has over 1.3 bn users. Since rollout started in 2010, the government has discovered and removed many barriers it initially created for users. As just one example - the need to provide fingerprints and iris scans to create an ID excludes many elderly people, manual labourers or those with certain disabilities.

While possession of an Aadhar ID is legally not mandated, the reality is that so many systems rely on it that not having an ID effectively causes exclusion and barriers. There are examples of officials manually intervening to ensure access for those who would otherwise have been unable to gain an ID. Recently, the need to increase access was one of the main reasons for introducing alternative biometrics - facial recognition, so that if either fingerprints or iris scanning does not work for a user, an identity can still be created in the system.

The above two examples are of a different scale, but clearly illustrate the need to build into your service the ability to deliver the same service in multiple ways to avoid exclusion. Offering different routes and direct support for those that need it most does not mean that the core service design is broken. It's simply the reality of public service provision at national scale.

You can help prevent unnecessary barriers by:

• Questioning when digital identity is really needed

• Baking in accessibility and the needs of marginalised groups into your service design and user research

• Allowing for change

• Ensuring your service design is open to change, and the use of alternative routes to verifying identity

• Making sure that non-digital user journeys are just as good and thought through as the digital journeys

These are all parts of solving this problem. The reality of providing a digital public service is you should expect that your teams will eventually spend just as much time eliminating the barriers it creates for the minority, as it will building the journey for the majority. Being prepared and open for this will help you keep your services inclusive.

To discuss digital identity with Public Digital, email anna.hirschfeld@public.digital

This post is part of a series on digital identity.