Selective Disclosure of Data: What It Is and How It Protects Privacy

Proving a condition does not always require sharing a complete document. Selective disclosure can reveal only certain data, but its protections depend on the system and subsequent use.

By Daniel Sardá · Published on

7 min read1,371 words

In this article · 8 sections

A service needs to check that you are above a certain age. You give it an image of your ID and, along with the answer it needs, it receives your name, date of birth, and other details it may not need. The check ends, but the information you handed over may remain in its systems.

Selective disclosure of data allows someone to share certain attributes of a credential while keeping others hidden. In privacy and digital identity, its purpose is to reduce the information a person reveals in each interaction. Here the term refers to that selection, not to selectively publishing data to support an argument.

Its value becomes clear with a question: if a service only needs to know whether you meet a condition, why should it receive the rest of your identity?

Sharing part of a credential without handing over the whole document

An attribute is a characteristic or piece of information: a name, date of birth, qualification, or membership in an organization. A credential can contain several attributes. Choosing which ones to present allows the information shared to fit a particular check.

Selecting information does not always require cryptography. With a digital credential, however, there is an additional challenge: the recipient must be able to verify the authenticity of what is presented without receiving every attribute. Simply covering fields in an image does not provide that assurance.

Selective disclosure mechanisms for verifiable credentials seek to address this problem. Their capabilities depend on the format and system used; a digital wallet does not automatically include every possible privacy feature.

Key idea: Selective disclosure reduces the data revealed in a presentation. Its authenticity and privacy protections depend on the mechanism used.

Who issues, presents, and verifies

It helps to distinguish three roles in the W3C's verifiable credentials model:

  • Issuer: the entity that issues the credential and stands behind its claims, such as a university certifying a degree.
  • Holder: the person who possesses and presents the credential, for example through a wallet application.
  • Verifier: the party that receives the presentation and checks whether it serves the requested purpose.

In a simple example, a university issues a credential containing a person's name, degree, and graduation date. If a service only needs to check the qualification, a compatible system might allow the holder to present that attribute without sharing the date.

The verifier checks the relevant cryptographic proofs and decides whether to trust the issuer for that claim. These are separate questions: verifying a signature does not by itself establish that the original information is true. An entity can sign incorrect information. Trust also depends on who issued it and how they checked the facts.

This architecture is part of digital identity, but it does not require every interaction to reveal a complete identity.

Proving age: two different approaches

Return to the service that asks whether someone is over 18. This is a hypothetical example: the threshold is not universal, and age verification is not necessary for every service.

If the check is justified, receiving a full date of birth may provide more information than necessary. In its statement on age assurance, the European Data Protection Board recommends limiting the result to what is necessary for the relevant threshold. This guidance from a European authority is useful when assessing the example's design.

Two approaches are worth distinguishing.

Presenting an already certified attribute

The issuer can include a claim such as “over 18.” If the system supports selective disclosure, the holder could present that claim while keeping other attributes hidden.

The recipient verifies a condition the issuer has already certified. It does not need to calculate the condition from a date of birth it receives. For this to work, the credential must contain the appropriate claim, and the recipient must be able to verify and accept it.

Proving a condition about hidden data

Another possibility is for a compatible mechanism to prove that a date of birth meets the requested condition without revealing the date. Some zero-knowledge techniques can prove claims of this kind without disclosing the underlying data.

This capability requires a specific design. Hiding the “date of birth” field does not by itself produce proof that someone is over 18.

Nor does all selective disclosure use zero knowledge. For example, the IETF's SD-JWT standard describes a selective disclosure construction using signatures, cryptographic hashes, and random values. It illustrates that there are different ways to reveal part of the information; their properties must be examined separately.

Key idea: Revealing an already certified “over 18” claim and proving a condition about a hidden date are different operations. Neither follows simply from removing a field.

Data minimization starts before choosing attributes

Data minimization means limiting the information requested and processed to what is necessary for a purpose. Selective disclosure can help achieve that goal during an exchange, but the decision starts earlier: what is requested, and why, must be justified.

A form that allows someone to hide their address but demands their name and date of birth without sufficient reason still presents a problem. Technically selecting fields does not fix an excessive request.

Important decisions also remain after the exchange: how long the recipient keeps the information, who can access it, and what it can be used for. Sharing less reduces the information exposed in that presentation; it does not prove that the recipient will delete it afterward.

That is why digital privacy depends both on the presentation mechanism and on the practices of those who receive the data.

Less exposure does not mean anonymity

Even when a presentation hides a name, it can leave elements that allow several transactions to be linked to the same person, including when that person's identity is unknown. This linking is called correlation.

The W3C warns of risks associated with identifiers and other elements of credentials and presentations. Metadata also matters: information about the interaction can provide context even when very little content is revealed.

The extent of the risk varies. The SD-JWT specification, for example, discusses limits on preventing presentations from being linked, including possible cooperation between issuer and verifier. That warning should be understood in the context of that mechanism, without attributing identical properties to every solution.

A serious assessment must ask whether transactions can be linked, what information each participant receives, and what happens during additional checks. The absence of a visible name is not enough to answer those questions.

How much control does the person retain?

From the perspective of individual liberty, being able to prove something without exposing information unrelated to the transaction can strengthen autonomy. It reduces what a public or private organization learns and lets a person tailor each presentation to its purpose.

That value depends on the conditions of use. A screen with checkboxes for selecting attributes offers little room to choose if refusing an excessive request prevents access to an essential service, or if no practical alternative exists. The technical ability to choose does not by itself establish that consent is freely given.

It is also worth asking who controls the infrastructure and what options are available to someone without a compatible device. These are criteria for judging a system, not guarantees that automatically come with a credential.

Key idea: Effective control depends on what you can withhold, what happens if you refuse, and how the recipient uses what you ultimately share.

Before accepting a request, consider five questions:

  • What does the service need to verify, and why?
  • Does it ask for a specific condition or for the full data from which that condition could be inferred?
  • Who will receive the information, and how long will they keep it?
  • Can this presentation be linked to other interactions?
  • What alternative exists if you cannot or do not want to use the system?

The age check makes the decision concrete: handing over an ID, presenting a certified claim, and proving a condition expose different amounts and kinds of information. Assessing selective disclosure means looking at those differences and following the data through to its subsequent use. That is where the extent of a person's control becomes clear.

About the author

Daniel Sardá is an SEO Specialist, a university-level technician in Foreign Trade from Universidad Simón Bolívar, and editor of Libertatis Venezuela. He writes on liberalism, political economy, institutions, propaganda and individual liberty from an independent, non-partisan perspective.

Read full profile