Skip to content

Latest commit

 

History

History
416 lines (329 loc) · 19.8 KB

functional-identity-primer.md

File metadata and controls

416 lines (329 loc) · 19.8 KB

A Primer on Functional Identity

By Joe Andrieu <[email protected]>

There are many ways to approach identity. For the Rebooting the Web of Trust workshop [1], we prefer a functional approach, focusing our conversation on how identity works and how it is used.

The varied facets of identity are rich. We inevitably bring our own hot buttons and agendas to any discussion of “what identity is”. Some engage from a philosophical perspective, others psychological. Some dive into political or cultural issues, while others dissect the meta-physical and spiritual. These different perspectives are valid views of identity’s impact on our lives. More than valid. Vital. They help answer the question of “Why?” Why identity matters, why we should care. Unfortunately, they also inflame passions and we sometimes talk past each other to make points that seem irrelevant to others, leaving people frustrated and unheard.

As engineers and system designers, we’re concerned with how things work. We want to fix what’s broken and build things that improve the world. To help us do that at Rebooting the Web of Trust, we discuss how things function. With identity, this functional perspective sidesteps seductive and inflammatory rabbit holes, without dismissing them. Functional Identity lets us investigate the HOW without prejudice to WHY, viewing identity systems based on how they work. Then, once we understand how they function, we have a solid foundation for discussion about how these systems affect individuals and society.

A Functional Definition

Identity is how we recognize, remember, and ultimately respond to specific people and things.

That’s it. We meet people and learn their names. We observe them and hear gossip and potentially consume related media. We remember what we learn. Then, we apply that knowledge to future dealings. Others do the same with us. Even our sense of our own identity is shaped by how we recognize, remember, and respond to our own actions and reactions.

Identity can be a wonderful thing. The joy of a child saying “Momma” or a lover calling out your name. The pride in your name on a diploma. The simple benefit of seeing another’s name tag at a workshop and making a note to yourself about a fascinating conversation. Identity enables so many benefits because it helps us keep track of people and things. It helps us recognize friends, families, and threats; it enables remembering birthdays, preferences, and histories; it gives us the ability to respond to each individual as their own unique person.

The functional approach reaches beyond digital systems to understand how identity works throughout society. Our identity is bigger than our digital selves. Our identities existed before and continue to exist independent of any digital representation. Digital identities are simply tools which help organizations and individuals manage real-world identity.

Unfortunately, digital systems can unwittingly compromise real-world identity. Sometimes this occurs because digital identity systems fail to consider external effects. Other times, it happens with systems that didn’t even realize they were dealing with identity information. A functional perspective allows engineers to see beyond static attributes and traditional notions of “Personally Identifiable Information” to better understand how engineering choices can impact identity, even outside their systems.

With a better understanding of how identity functions, we will be able to build systems that enhance privacy and human dignity, while improving identity assurance and security.

Identity Systems

An identity system is a collection of tools and techniques used to keep track of people and things.

As individuals, we do this naturally, in our minds. We name things, then use names and distinguishing features to remember what we learn. We treat people differently based on their identity: treating our friends and family differently from strangers and known threats.

Organizations create processes, software, and services to achieve similar ends. These identity systems are best understood in terms of how they function, which is the same as how identity has worked since the dawn of mankind.

For some, identity systems are provocative because they help organizations keep track of people. They trigger associations with Big Brother and the surveillance state, inspiring dystopic visions of embedded chips and tattooed serial numbers. Discussing a ubiquitous “identity” layer for the Internet conjures fears of government and corporations constantly tracking what we do, online and off.

It should. Because, in fact, these are legitimate abuses of identity feared by civil libertarians and freedom-minded people everywhere. When we talk about identity systems, we are necessarily talking about how we keep track of people and things. Do it badly and we risk accidentally building our own Panopticon prison [1]. Fortunately, by understanding how identity functions, we can avoid, mitigate, and minimize such abuses.

Terminology

Identity is an information processing task that maps persons or things we might know to what we actually do know. This mapping may be entirely digital or it may rely on physical devices or sensors to provide input from the “real” world. Functional Identity focuses on the innate information processing that occurs in every identity system whether it is natural or engineered, purely digital or physically grounded.

In the domain of information processing, informational assets and processes are the essential nouns and verbs that describe how any system works. This next section presents a concise and complete set of nouns and verbs that fully describe identity systems. When you understand these nouns and verbs for a given system, you will understand how that system creates, uses, and impacts identity, both within its boundaries and in the broader context of individuals lives in society.

In the diagrams below, the blue boxes are nouns and the red ovals are verbs. Together they comprise the building blocks for describing identity systems.

We start with the simplest identity system, using three nouns and a verb:

Subjects are entities—people or things—under consideration.

Identifiers are labels which refer to entities. They are used to keep track of what we know about those entities.

Attributes are what we know about people and things. They describe the state, appearance, or other qualities of an entity.

Correlate means to associate attributes with particular entities, to associate what we know about someone with either an identifier in the system or a subject in question.

Identity systems correlate subjects with attributes in two ways. First, attributes are associated with identifiers referring to specific subjects, thus building a body of knowledge. Then, when we recognize a subject, we associate them with one or more identifiers, and in doing so, associate them with everything associated with those identifiers.

In digital systems, this set of related attributes is sometimes referred as a digital identity or profile.

Input and Effect

We learn or acquire identity information over time, then apply what we’ve learned to various interactions, usually elsewhere.

Acquire** means to gather identity information for use by the system.

Apply means to use identity information to affect change outside the identity system, typically to moderate an interaction of the subject with a related system.

Identity information might be acquired by observation or by importing from elsewhere. We may learn about someone by watching them, or we may learn through references, rumors, and reputation. Identity systems acquire new information throughout their operational life, just as we continue to learn about people throughout our lives.

Once acquired, identity information must be applied in a specific situation to have impact. If we know something about someone and that information never influences our behavior and is never shared, it doesn’t affect the world. The way that identity information is applied tells us how an identity system affects our lives.

For example, a website might apply the email associated with my account to allow me to reset my password or it may send me unwanted advertisements. The U.S. Transportation Security Administration (TSA) applies the information on its no-fly list to prevent those identified as potential threats from flying.

Making New Ideas

We gain new insights by considering both existing identity information and previously unrelated observations. Identity is more than just what we know about people and apply to our interactions. It’s also how we make judgments based on what we know, gaining insights into character, capabilities, and proclivities.

Raw data** are data which may or may not contain information relatable to a person or thing.

Context tells you why you can rely on any given identity information and what you may do with it.

Reason means to evaluate existing identity information to generate new derived attributes.

New attributes are created by reasoning using raw data and known attributes. By applying reasoning to existing observations and related knowledge, we can gain insights that neither the subject nor the original author anticipated. Raw data such as search history, web browsing, and the time & location information captured by our phones, may contain identity-related information, even when that was neither the purpose nor the intention at the time of capture.

The contexts associated with identity information inform us about appropriate use, including the evidence needed to understand how trustworthy it is. Context answers questions such as:

  • Where did it come from?

  • How did we get it?

  • When was it created or modified? By Whom?

  • What purposes, privileges, and responsibilities are attached?

In short, context allows you to evaluate if a given piece of information is credible.

In many real-world identity systems, like that enabled by state-issued driver’s licenses, the context is implicit, spatial, and temporal. Online identity systems lack this physical immediacy and need to use other mechanisms to capture and understand context.

We also reason using known attributes to derive new ones. For example, we calculate a person’s age based on the birthdate on their driver’s license to determine if they are old enough to drink legally. Credit companies evaluate recent income, past transactions, and projections of future income to set interest rates and make loan approvals. We remember how people treat us and alter our behavior in future interactions. If someone repeatedly breaks their word, we may stop depending on them.

Governing Identity Information

We go to great lengths to manage identity information.

Govern means to manage the creation and

flow of identity information so the right people have access for the right reasons at the right time.

Sometimes we keep secrets to prevent information from reaching certain people. We do this with tools like encryption, access control, and minimal disclosure. Legal agreements between people, businesses, institutions, and governments specify appropriate use of certain information while laws, regulations, and the courts allow governments and institutions to oversee, monitor, and intervene in the capture and use of identity information. How identity systems govern who controls certain information defines how they preserve and respect privacy.

The right to keep private information private is often referred to as the right of privacy. Many people feel their privacy is threatened because so much information is shared over the Internet, in our workplaces, and through our devices. Information we share in different contexts (business, family, community, etc.) can leak unexpectedly and undesirably into other contexts.

It is very difficult, as individuals, to track of all the ways we are publicly or privately tracked. Information is shared on social media, tracked in Internet searches, monitored when using navigation software, and captured as we use our phones. The sheer magnitude and complexity of information sharing means the average person is essentially incapable of making informed decisions to consent to appropriate use. Some people give up, divulging personal information without regard to consequences. Others opt-out, participating as little as possible in our digitally connected world. Governance defines who gets to control this complexity and how we do so.

Bridging the Gap

The nouns and verbs above are grounded in the world of technology and may be unfamiliar for the average individual. More conversational synonyms are presented in the tables below. Use the most appropriate terms for your audience.

People, Places and Things

This is the point of identity: those people, places, and things we recognize.

Technologists Laypeople Common meaning
Subject Person or Thing Someone or something under consideration. The focus of interaction or discussion.

Identity Information

These are the abstract nouns of identity, the informational assets created and used by identity systems.

Technologists Laypeople Common meaning
Identifiers Names Refers to entities. Used to keep track of people and things.
Attributes Knowledge What we know about people and things. How we describe the state, appearance, or other qualities of an entity.
Raw data Observations Data which may or may not contain correlatable information.
Contexts Situations Information which allows us to evaluate if another piece of information is dependable.

Identity Actions

These are the verbs of identity. These are the actions taken by identity systems working with identity information.

Technologists Laypeople Common meaning
Acquire Collect Intake or generate identity information for use by the system.
Correlate Relate Associate attributes or observations with particular entities. We associate what we know about someone with either an identifier in the system or with a subject in question.
Reason Reason Evaluate existing identity information to generate new beliefs, expressed in attributes, captured in statements.
Apply Apply Use identity information in a system, typically to moderate interactions with known entities.
Govern Control Manage the creation and flow of identity information to the right people at the right time.

For technologists: we assign identifiers to subjects. We collect raw data and correlate attributes to the subjects we track, in specific contexts. We reason over raw data and attributes, to derive new attributes. We then apply this information to current and future interactions with subjects. We govern identity information to preserve privacy and give appropriate controls to the right parties.

In more ordinary language: we give names to people. We collect observations and linking those observations to people, remembering knowledge about them. We reason over these observations and knowledge to generate new knowledge. We then apply what we know when dealing with those we recognize. We control identity information to preserve privacy and to protect those we love.

This is the vocabulary of Functional Identity, a way to discuss and understand identity in terms of functionality: how it works and what it does for us. This is the language of identity for Rebooting the Web of Trust.

Why?

Engineers, entrepreneurs, and financiers have asked “Why are we spending so much time with a definition of identity? Why not just build something and fix it if it is broken?” The vital, simple reason is human dignity.

When we build interconnected systems without a core understanding of identity, we risk inadvertently compromising human dignity. We risk accidentally building systems that deny self-expression, place individuals in harm’s way, and unintentionally oppress those most in need of self-determination.

There are times when the needs of security outweigh the desire for human dignity. Fine. It’s the job of our political systems—local, national, and international—to minimize abuse and to establish boundaries and practices that respect basic human rights.

But when engineers unwittingly compromise the ability of individuals to self-express their identity, when we expose personal information in unexpected ways, when our systems deny basic services because of a flawed understanding of identity, that’s avoidable tragedy. What might seem a minor technicality in one conversation could lead to the loss of privacy, liberty, or even life for an individual whose identity is unintentionally compromised.

That’s why it pays to understand identity, so the systems we build intentionally enable human dignity instead of accidentally destroy it.

Summary

Functional Identity focuses on how identity works. At the Rebooting the Web of Trust, we ground our work in the functional notion of identity and avoid the psychological, cultural, political, and philosophical. These notions are important, but they can also distract us from understanding the technical choices involved in building and using identity in today’s networked world.

This functional notion of identity began with a conversation at the Internet Identity Workshop [3] in May of 2016, followed by conversations at ID2020 [4] and the second Rebooting the Web of Trust workshop that summer, resulting in the paper “Identity Crisis” [5]. It continued in subsequent meetings in all three venues, and in two articles published by the People Centered Internet [6, 7].

This primer represents a current take on that conversation, geared to help Rebooting the Web of Trust participants communicate more clearly and collaborate more effectively. We encourage your feedback and look forward to continuing the conversation.

[1] http://weboftrust.info

[2] https://en.wikipedia.org/wiki/Panopticon

[3] https://iiw.idcommons.net

[4] http://id2020summit.org/

[5] https://bitly.com/IdentityCrisisPaper

[6] https://peoplecentered.net/2017/06/11/speaking-of-identity/

[7] https://peoplecentered.net/2017/07/26/how-identity-can-enable-a-people-centered-internet/