Archives for January 2018

What is Self-Sovereign Identity?

Kaliya Young · January 23, 2018 ·

Here is me on a the Creative Futurism podcast talking about Self-Sovereign Identity. We don’t go into the technology details but stay with the big picture concepts. What is happening is we finally have the pieces coming together for an identity layer of the internet where people are in control of their own identities.
This technology will have significant implications for how people interact with governments, how people manage their relationship with businesses, how we do banking, how we manage our medical and educational records and how we relate to each other peer to peer.

Self-Sovereign Identity is now possible because of these technologies.

Probably the most important one, that is often not recognized or observed that often because we focus so much on technical technologies rather then the social technologies that makes their innovation possible. I just wrote an article about this for Open Democracy.
The Social Human Fabric: This is critical to why we actually are having these breakthroughs woven together to create the open standards for self-sovereign identity to become real. A dedicated group of individuals that have been actively meeting face to face twice a year for over 12 years at the Internet Identity Workshop, and other industry events like the RSA Conference, Cloud Identity Summit and the new kid on the collaboration block Rebooting the Web of Trust.
Smart Phones: The iPhone is 10 years old and is key to supporting the individual having a computer in their pocket to manage their identity attributes and private keys.
Cloud Computing: Key for individuals having cloud agent and wallet
PKI – Public Key Infrastructure: This is not new either first created over 25 years ago. It is the basis of encrypted communication but there have been huge usability challenges that have prevented usage
Shared Ledger Technologies (also called Distributed Ledgers or Blockchains): Key for storing Decentralized Identifiers and the public keys associated with them. This creates the framework for decentralized yet globally resolvable
Open Standards for Decentralized Identifiers – DIDs: Community leadership worked hard to support those working on decentralized identifiers on block chains to actually collaborate on a minimally common set of protocols.
PairWise or Directed Identifiers: For each relationship with each person or entity in the system individuals create a pairwise identifier – a unique identifier a new DID just for that relationship. This means they don’t leak information by using the same DID for different connections / relationships.
Open Standards for Verified Claims: This work has been going on for many years at the W3C and leverages the Decentralized Identifier standard to empower issuers to issue verified claims to individuals. These include courses taken or degrees issues

How does it work?

This is still being figured out.
I have seen working wallets in the lab.
Individuals have to get an Edge Wallet – an application on their smart phone and set up a relationship with a service provider to support their cloud wallet. Wait a second they have to trust another service? Yep think of this like a bank account – we have a network of service providers that help us store and manage our money – but it is our money – not theirs. We always have the right to move service providers to change who we trust to work on our behalf in this ecosystem. Then using these tools we create a Decentralized Identifier a DID and this gets published into a distributed ledger. Its a really really long number that we prove we own because attached to it in the DDO (DID Document) is a public key. In your wallet is a private key that proves you are the owner of the public key in the DDO and thus the owner of that identifier.
All identity information is stored in the cloud agent. NO IDENTITY INFORMATION IS EVER STORED on CHAIN.
Individuals who want to maintain separation between all their different identities will be able to do so because each relationship has a different DID associated with it.

What can it Do?

This is the infrastructure for individuals to start sharing and collecting identity information that they collect and manage without third parties seeing where they share it/how they transact.
It is also the infrastructure for individuals to start connecting to each other one-on-one to connect with secure encrypted channels for sharing and connection.
It is new infrastructure for social sharing and collaboration providing a place to root individuals identities that provide the opportunity for real alternatives to centralized networks like Facebook.

Identifiers: A Field Guide

Kaliya Young · January 4, 2018 ·

Tim Bouma wrote a post about Trusted Digital Identity.
In it he unpacks several terms including this one

Identifier: Anything (name, numbers, symbols, etc.) that uniquely distinguishes a member of a population from another member.

I don’t think this definition gets at the complexity of all the things that can be identifiers and how to distinguish them from one another. In 2014 I began working on what I called a field guide to a whole bunch of aspects of identity. This is the section that I wrote about identifiers. I am republishing it here and would love feedback – which I will incorporate into this version of the post. A complementary post is up that looks at a very nuanced discussion of what an identifier actually is What is an Identifier?.
I should say that the frame of a field guide is intentional. We are in a landscape of a range of identifier types – that we need to understand and distinguish among just like the autobahn society created the field guide to birds so we could understand them and their characteristics in different ecosystems.

Identifiers

For people Names are a special class of Identifiers. They are both self-asserted by people and are used to refer to them and acknowledge them in social context.

System Identifiers

In systems, bureaucratic, digital and techno-bureaucratic identifiers are alpha numeric string pointers at/for people in systems.
This may seem simple but their are many different types and a person with a record in a system will likely have more then one type. To get these different types of identifiers I will share different examples.

Persistent Correlateable Identifiers

This type of identifier is re-used over time within contexts and across multiple contexts.
Examples
Student Number – When I enrolled at my university I was assigned an 8 digit student number. This number was persistent over my time as a student at the school. When interacting with school institutions I was asked to share this number so that activity could be linked together across different facets of the institution.

Social Security Number – This number is issued by the federal government to those born in the US as part of the standard process for being born. It is meant to help those who submit money to the SSN system and when they retire be able to collect money from the system.
Aadhar Number – This created by Indian government for each resident of India. To get a number and individual enrolls all 10 finger prints, two iris scans and a photograph – 13 biometrics. Their is a check to see that this person has not registered already and then a number is issued to them.
Phone Number – People today often have a personal number that they use across many different contexts. It is common place to ask for a phone number to be able to contact a person. What people don’t know is that those are used to look people up in data broker services. The phone number is used to link together activity across contexts.
E-mail Address – Many people have one personal address and use it These are often used across different contexts. What people don’t know is that those are used to look people up in 9data broker services like RapLeaf.

Assigned Identifiers

These are assigned to individuals by systems of government and businesses to support them being re-identified when they return to an entity to interact again. This is from section 7.4.2 of the PCIM Validation Standard.

Once associated with a person, an assigned identifier uniquely distinguishes that person from all other persons in a population without the use of any other identity attributes. Examples of assigned identifiers include birth registration numbers, driver’s license numbers, and social insurance numbers. The following considerations apply to the use of assigned identifiers:

Assigned identifiers may be kept internal to the program that maintains them. Examples of internal assigned identifiers are database unique keys and globally unique identifiers.

Assigned identifiers maintained by one program may be provided to other programs so that those programs can also use the assigned identifier to distinguish between different persons within their program/service population; however, there may be restrictions on this practice due to privacy considerations or legislation.

Certain assigned identifiers may be subject to legal and policy restrictions. For example, the Government of Canada imposes restrictions on the collection, use, retention, disclosure, and disposal of the social insurance number.

Directed Identifiers

A directed identifier is created to support individuals using different identifiers in different contexts. The purpose of this is to inhibit the ability to link records across contexts.
Examples
The British Columbia eID System – This system enrolls citizens and issues a card to them. When the card is used to access different government systems by the citizens. It does not use one identifier for the citizen. Rather for each system it uses a different identifier for the system – an identifier directed for a particular system.

Decentralized Identifiers, DIDs – this type of identifier can be easily created and therefore can be directed – meaning that individuals only use a particular DID for interacting with a particular institution. The reverse is also true. Institutions can created a separate DID for each connection they have to each individual. Sovrin and Veres One are working on creating ways for directed identifiers being accessible to individuals and institutions to manage their connections to one another.

Defacto Identifiers

By combining a name names, and key attributes together systems use this combination to create a defacto identifier which uniquely identifies a person often in the context of a whole society. An example is the us of “name” “birth date” and “birth place”. It seems innocent enough to be asked for one’s name, birthdate and place but this becomes a persistent correlateable identifier to link and track activity across many systems. The creation of defacto identifiers that are persistent and correlateable limits people’s ability to control how they present in different contexts.

In the process of creating a feedback loop related to this article Tim pointed to section 7 of this work under development by thePan Canadian Identity Management efforts. Now his definition above makes more sense. In section 7.4 they talk about identity this way. They are oriented to defect identifiers to ensure uniqueness.

A property or characteristic associated with an identifiable person is referred to as an identity attribute or an identity data element. Examples of identity attributes include name, date of birth, and sex. For any given program or service, identity information is the set of identity attributes that is both:

Sufficient to distinguish between different persons within the program/service population (i.e. achieve the uniqueness requirement for identity); and

Sufficient to describe the person as required by the program or service.

Section 7.4.1 says this:

The set of identity attributes that is used to uniquely distinguish a unique and particular person within a program/service population is referred to as an identifier

Opaque Identifiers

An opaque identifier is one that does not give away information about the subject it identifies.
Examples of Opaque Identifiers
The BC Government eID program has at its core an opaque identifier on each card – it points to their card record. It is just a number with no meaning. If they loose their card a new opaque identifier is issued for their next card.
Examples of Non-Opaque Identifiers

Examples of Non-Opaque Identifiers

National Identity Number in South Africa contains a lot of information it is a 13-digit number containing only numeric characters, and no whitespace, punctuation, or alpha characters. It is defined as YYMMDDSSSSCAZ:

YYMMDD represents the date of birth (DoB);
SSSS is a sequence number registered with the same birth date (where females are assigned sequential numbers in the range 0000 to 4999 and males from 5000 to 9999);
C is the citizenship with 0 if the person is a SA citizen, 1 if the person is a permanent resident;
A is 8 or 9. Prior to 1994 this number was used to indicate the holder’s race;

• Z is a checksum digit.
The US Social Security Number is created via a formula and so the number gives away information about the person it identifiers.
Phone numbers give away information about the metro region that a person was issued the number from.

End-Point

Some identifiers that represent people are also end-points to which messages can be sent.

Physical Address

It is often forgotten in conversations about digital identity that we had a system of end-points for people before networks known as a mailing address. They system of mailing addresses was developed and is maintained by the US postal service.

Network Address

Phone Number – Now with cellular phones people have their own phone numbers (not just one for a household or their workplace as a whole). This permits both voice calls being made, text messages and MMS Multi-Media messages. The name space for phone number originates from the ITU-T. They are globally unique. They are also recyclable.
E-mail Address – These addresses permit people to send messages to the address they have. They are globally unique. The name space for domain names resides with ICANN. They are also recyclable.

Device Identifier

Many digital devices have unique identifiers. Activity on digital networks can be linked together by tracking these activity originating from particular devices even if people using them .

Non-End-Point

These are identifiers that do not resolve in digital or physical networks.

Document Identifiers

Documents like birth certificates have serial numbers that identify the document.

Document Validation Systems

These systems are used to look up which documents are infact valid. When properly constructed they don’t give away any information about the person. Those using the system type in the serial number of the document and information it contains and the system simply returns a Yes/No answer about weather it is valid or not.

Beacons

A beacon actually broadcasts from a digital device a persistent correlateable identifier to any device that asks for it. It creates a form of tracking people and their devices in the physical world.
Examples
RFID chips, cellular phones, laptop computers

Polymorphic

These systems generate different identifiers depending on context.
Examples
The BC eID system way of using one card that then supports the use of different identifiers depending on context.

Time Limited & Revocable

Some identifiers are created and point at a person but are revocable. An example is a phone number that is after one stops paying one’s phone bill for a month is re-assigned to another person. An employee at a company may have an employee number that is revoked (no longer valid) once employment is terminated. A passport number is an identifier that has a time limit it is good for 5 or 10 years. A landed immigrant card (green card) in the US is only good for 10 years.

Un-Revocable

These identifiers are persistent and are not revoked. Examples include Social Security Numbers.

Identifier Issues

Identifier Recycling

Some identifiers are in systems where identifiers that point at one person can be discontinued (they stop paying their phone bill or using their e-mail address) and then the identifier can be re-assigned to a different user.

Delegation (Acting on Behalf of Another)

This functionality is critically to a variety of user populations. Elders who want to delegate access to their accounts children. Service professionals who have contractual relationships with clients such as an accountant managing access to financial & tax records. Most systems are designed with an assumption that people themselves are the only one accessing accounts. This creates a problem when people want to delegate access they have to turn over their own credentials so the person they are delegating to “pretends” to be the actual user.

Stewardship (Care-Taking – Oversight)

Their is another role that is slightly different then delegation when someone turns over a power of attorney like function for a particular account/set of functions. Stewardship of identity is the type of relationship a parent has for a child’s identity or the type of care needed to help the mentally disabled with their interactions online.

The Mesh of Pointers

We end-up with a way that identifiers work together as a web of pointers towards a particular individual.

What is an Identifier?

Kaliya Young · January 4, 2018 ·

Tim Bouma wrote a post about Trusted Digital Identity.
In it he unpacks several terms including this one

Identifier: Anything (name, numbers, symbols, etc.) that uniquely distinguishes a member of a population from another member.

Part of Tim’s definition resonates with a conversation I had with Jean Russel in 2009 that I thought I would share that to support a very nuanced and specific conversation about what identifier actually are and how they work in the physical world and digital world. This complements the Identifiers: A Field Guide post.
Identifiers
Kaliya and Jean Russell share a dialogue, learning from each other about reputations and currency. (I write in third person because I want to attribute appropriately to each, and yet this is done together). We have a sense of the overall map of ideas, and we want to start with some core concepts that the work depends upon.
We begin with identifiers. We discuss below what identifiers are and how they work in meat-space. Our next post covers identifiers in the digital context.
Jean: SO….What is an identifier?
Kaliya: An identifier is a pointer to a person or an object
Jean: A pointer to a person or an object?
Kaliya: There are generic identifiers – rose, cup, chair…
Jean: So a word can be an identifier?

Kaliya: Yes. To have a more specific identifier “the green chair over in the corner” identifies it (the specific green chair) …relative to others in the same context – a room, for example.
Jean: Okay, I think I get what you mean by pointer. An identifier allows you to identify something to someone else in a shared context.
Kaliya: Yes. So people’s names identify them in our shared social spaces. They are identifiers too.
Jean: So in meat-space we are using identifiers all the time when we use language together.
Kaliya: However, I am not my name, I have a name – it points to me. You have a name – it points to you.
Jean: Okay, so the name and what it refers to are not the same thing. One is pointing at the other. And there are different kinds of identifiers, then? Like chair is vague and green chair in the corner is specific and my name is specific to me, pretty much.
Kaliya: Chair is a generic identifier, yes. Well, it is specific to you in a social context. Green chair in the corner is more specific. I might want to identify a very particular green chair. I would look on the chair to find the manufacture serial number for it, or I might want it in my company/personal inventory and “assign” it a number identifier for that specific chair.
Jean: Right, so there are degrees of specificity in identifiers.
Kaliya: So people’s name are specific in a social context. They might be more or less “specific” because there is more than one person named Jean in the world and even with my name there is more then one Kaliya. But in my social world – the people I know – I am the only Kaliya. I know several Mary’s though so I have to get more specific when talking about them using a last initial or a last name.
Jean: Okay, so there is an element of uniqueness that is important in an identifier? To successfully identify the object, the identifier needs to be unique?
Kaliya: Yes, unique within the context.
Jean: So we seem to navigate this pretty well in our everyday lives, and we ask for more specificity when we need it.
Kaliya: Yes.
Identifiers in a Digital Context
Jean: Can you explain what identifiers mean in the digital environment?
Kaliya: So, when I am at a dinner party with two Mary’s and having a conversation you signal who you are talking to/about via gestures and stuff – you layer in more info about who you are talking about. Or you might, in a conversation in digital chat, say “Mary R” or “Mary H” because you don’t have bodies and social gestures to layer in. So when we go into digital realm – on the internet, what is the context we are in. So when someone goes to a website and gets an account, they get a username.
Jean: Right, I do that all the time. What does that mean?
Kaliya: The site – often checks to see if anyone else has that username, if they do…you can’t have it because someone else has “that” identifier already.
Jean: So I might be able to get ‘Jean Russell’ on one site but not on another, for example?
Kaliya: Well likely you wouldn’t have a space in your username, so ‘JeanRussell’ or ‘Jean_Russell’
Jean: Ok, so no space, so the code can read it, but I might get ‘JeanRussell’ on one space but not on another, on that next space I get ‘JeanRussell6′
Kaliya: In a way, identifiers for people are like digital bodies, but they were weird cause they wouldn’t let you bring a “body” from another site/context into their site/context.
Jean: Every site you went to – every new site – they would make you get a new “body” a new identifier for that site. Ah… I don’t want to keep track of all those bodies. This is so annoying. I am one person. I want my name to be the same regardless of what site I am on.
Kaliya: Well yes – exactly, so the question is how do you have a unique identifier, that “works” for you across the whole internet. This is what OpenID does. It creates a way for you prove you “own” or “have control of” (as in knowing the password for an account). You need to be Unique within a bigger context then just that website, so the large sites allow users to take the identifier within their space and use it other places. So you can use your Yahoo! ID or MySpace ID and log into other websites. OR you could go and buy a domain name just for you – and use it. So I own http://www.kaliya.net and it is set up so that I can use it as my open ID.
Jean: Well that seems to make it easier. But I still don’t get how it is working compared to the JeanRussell who already signed into this site I am trying to get into
Kaliya: You are just JeanRussell within that context – that website. Identifiers in the digital world, to be effective – need to be unique globally. URLs are all Unique. There is a name space….and domain names – are unique, a global registry, makes sure that no two people/companies/organizations own the same domain name.

Digital Bodies and User-Centric Identity
Jean: Kaliya, we left off our last chat talking about digital bodies and the importance of context with identifiers. Can you say more about digital bodies?
Kaliya: Well lets start with physical bodies – we have just one of these. So when we walk around in physical space people recognize us because we are in the same body we were last time. We only get one and over time is ages but basically it doesn’t fundamentally change and we can’t “get another one.” Last time we talked about identifiers and having the ability to have a globally unique one that you could take with you around the web. This gives you a freedom to move between websites and take your “digital body” with you. The difference is that in digital space you could make yourself several different “digital bodies or identifiers” that were globally unique that you would use in different contexts.
Jean: I am already a second body by creating the first digital body, right? Since it isn’t my physical body?
Kaliya: 🙂
Jean: So having many bodies is even more to keep track of and create?
Kaliya: The digital identifier you create that points at you – is like another digital body. Maybe you want to just be http://www.jeanrussell.com everywhere on the web. Maybe you want to have a professional life “identifier” and a personal life “identifier” that separates those two aspects.
Jean: You mean I can manage those bodies instead of having each platform define them for me?
Kaliya: Yes. An example that was brought up yesterday here at Super Nova by danah boyd was that of a teacher. That a teacher is working in front of children – they can’t be seen to be sexual (having a normal dating life) or drinking alcohol (as a normal social adult). So this is an example where someone in that profession would create an identifier they use to connect to their students on social networks and comment on blogs etc.
Jean: Right. That makes sense. Even in my physical body in the analog world, I am showing different facets of myself in different contexts.
Kaliya: They need to have a different identifier they use for their social connections to other adults – in their dating/social life. That same teacher might be politically active – as they have a right as a private citizen to be and those political views well within the spectrum of points of view that are acceptable might not be “the same” as those in their particular town or neighborhood – say a strong environmentalist in a very coal producing town. So they want to take action and voice opinions and share with others who are other active citizens. They would need a different digital identifier for that.
Jean: So, it feels like an advantage to have the ability to manage these digital bodies based on the context they show up in? And thus the community they mesh with in that context?
Kaliya: Back to our first conversation it would be great if they didn’t have to get a new identifier each time they went to a different environmental site – a portable one for them within that context of environmental activists. Yes, contextual management is important. The tools to support individuals doing this are just beginning to be conceptualized and developed.
Jean: So what I hear you saying Kaliya, is that we need our digital bodies to be a reflection of the facets of ourselves and the intersection of those facets with the communities we participate in. This is not defined by platform as much as it is our practices online.
Kaliya: Yes – an we need open standards that give us the freedom to move around the web with identifiers (digital bodies) from one website to another. This has to do with the underlying architecture of the social web that platforms build on. How we use these platforms and tools is complex. To have good practices, we need development of “web” (which had internet below it) and then on top of that is a layer where identifiers are – and applications that use them. then there is an emerging set of standards to move information we generate in social contexts around between sites these are called activity streams. So a website is a context, a group within a site is a context too. Each google group you are in is different – its own cluster of people.
Jean: Right, although there might be some overlaps, that can’t be assumed that I want to show the same facet of myself to all of my google groups.
Kaliya: There are sort of meta contexts – so a network of environmental activist sites would be an example of that.
Jean: Right, a site like Zanby does that for One Sky. Or Ning, or is that more of a tech context and not a purpose context. [Kaliya: and neither uses OpenID]
Kaliya: I guess you can think of it as topic contexts and platform contexts. One of the issues is that most platform contexts do not support being able to switch between different login/handles/identifiers very easily at all. You might have a personal yahoo account and a professional one, on ning too, same deal.
Jean: Right, like on twitter, I was working around that by using api clients or using different browsers!
Kaliya: Right, or logging in and logging out. Mozilla is working on a project to help people manage their ID’s within the browser. The platforms would like us all to “just have one identity” and not switch between but this is not realistic.
Jean: Right, getting back to that teacher example – she may want to be in touch with students on facebook… and want to keep her personal life in a different name there. So we have a social practice for doing that, but the tools don’t yet adequately reflect that.