Keynote: The Role of Counselors and Applied Identity

Posted

[I’ve been kicking around this notion of counselors for a few years now. The first time I debuted the idea it was at the OpenID Foundation Japan conference in 2019. I have been refining it over time. The refinement continued through keynote season this year. What follows is both a recording of my delivery at EIC 2024 as well as the text. You’ll notice that the two differ… and that is a result of my process. Where I start a talk and what gets on stage tends to differ… but in this case they differ reasonably significantly. Truth be told, I didn’t love this talk, but I got a lot of positive response to the actual delivery… which goes to show that what you, the speaker, things of a talk, and what the audience thinks can be very different! Watch the video or read the text or both and, regardless, enjoy!

— ian 20 June 2024]

Delivery at EIC 2024

The Bad News Up Front

What if I told you, despite decades of work, we, myself very much included, as an industry, haven’t made much progress? What if I told you that we have barely begun to actually solve the problems set before us? There’s a part of you that wouldn’t believe me.

That part of you responds with statements like: look at all these standards; look at all these products; look at all the money that has flown into IAM from the venture capital and public markets. And these statements are, obviously, correct. We have built and rebuilt a treasure trove of standards. This work is hard work. This work is very necessary work. And we’ve built a wide range of products with the standards. We have done as well as we could at the time we did. And over the last few decades it is very true that billions of dollars have been spent in the process. 

But… the problems we are meant to solve still remain. We still struggle with user account management and access governance. We still struggle with bridging the gaps between what we do and our security and developer peers.

The better half of the First Family of Identity, Pam Dingle, once stated something to the effect of, “We haven’t solved anything yet. We are just now building the tools we need to go solve something.” She was and continues to be spot on.

Simply stated, we struggle with applying the tools we have built. We struggle to use IAM to solve the problems our enterprises, institutions, and our citizens face. We struggle to APPLY identity and access management.

Consider that no one (other than people similar to those in this room) actually cares about signing into stuff. Just as no one cares about a picture on their phone that purports to be a driver’s license. What they do care about is what they can actually accomplish after signing in or showing their mobile drivers license.

But is exactly what the focus of identity management ought to be. We need to apply identity management to help achieve the goals that people really care about. We need  applied identity management… the modern era calls for AIM.

And I believe that we are now closer than ever to being able to apply identity and access management to actually achieve something – to do something. And the way I propose we do this is via the concept of: counselors.

Citing sources

The work we do doesn’t happen in a vacuum. We do hard work together – in our enterprises, in standards bodies, in countless online meetings and conference calls, in professional communities, and yes, even in cocktail bars in Berlin/Vegas. Before I go too far into this topic of counselors, I want to thank people who have contributed to this concept for MANY years. Many of whom have never heard me talk about but all of whom have contributed in their own particular ways.

I want to thank

  • Mike N and Bob B for Limited Liability Personas
  • Nat Sakimura for privacy notices
  • Andi Hindle to editorial and inspirational guidance
  • Andrew Nash for his 2018 Identiverse talk on identity management vs account management
  • Pam Dingle for the previously mentioned observation and host of other things

All of these people help contribute to this talk through years of interactions, turns of phrase, and their own hard work.

On the shoulders’ of these giants I stand. So let’s see what I can build with such a foundation!

What is a Counselor?

I am using counselor to mean a digital agent that can act on your behalf, give you advice, make suggestions, and look after your interests. It is an active client on all your devices that can advise you and protect you. They can step in before you share data with a service that could be considered risky. Imagine something stepping in before you hand over your form of payment and email address and suggest using an anonymous one instead. Imagine that service can even generate anonymous forms of payment and pseudonymous email for you. That is the future role of a counselor.

I want to be able to say to my counselor “find me an airplane to Berlin in premier economy for these dates in June.” The counselor in turn should know I will be leaving from Washington DC, that I am a Star Alliance flier, and that I like window seats. And with that the counselor should be able to go off and complete the entire transaction. 

In short a counselor uses your digital identities on your behalf to actually get something done. It is the embodiment of applied identity management.

Critical components of a counselor

There are 6 critical components to counselors. Let’s start with the most familiar. [use building diagram with different shapes to indicate whether we have such a thing, in development, or just don’t have yet]

Credential Broker

In order for a counselor to do something on your behalf it has to act, from a digital identity perspective, as one of your personas. And that means presenting your credentials. Good news is we have these in mass markets. Between password managers and passkeys powered by secure enclaves, now more than ever people can store and manage where and how they sign into stuff. And one can see how, if verified credentials and their associated wallets come into mainstream, they too fit into this pattern. But the important point is that we need a broker over the top of these because no one today or in the next decade will ONLY have username/passwords or only passkeys or only verified credentials. We will all have all three in differing amounts, and something that knows what you use where and how to securely present it, a la this credential broker, is required.

Data Manager

But as we know life is more than just signing in… right? We remember that life is more than just signing in? [do an audience participation thing here – make them say it.] If we really want to get something done, we need to provide and gather information… like shipping address, form of payment, proof of residency, student ID number, etc. Now to some degree we have these too. Form filling is, at least to some degree, a common part of password managers. Since not every use case is a high assurance use case, password managers work adequately. Let’s face it there are many scenarios in which the service neither needs nor cares to verify information provided by the individual. Not every use case is rooted in a regulated sector; not every use case requires KYC; not every needs for documentation of self-hood. May this never change. 

But there most definitely is a need for higher assurance and more, shall we say, formal verifiable data. But that’s not enough. Some of this data needs to be verifiable by the party to whom we are presenting it.  Hopefully the work on verified credentials and wallets will provide a standardized (or at least standardizable) means for data format and presentation that can span both use cases that require verification and those that do not. If not, we’ll always have form fillers for data presentation and have CSV for format.

So far what I describe exists today. Are we still working here? Yes, of course. But we have mass market adoptable and adopted things that actually work today.

The Noodge

There are many dark corners of the internet. And while many of us have developed an inherent sense for danger in the real world such as “don’t go down that alley; it looks sus” or “don’t take money out of that cash machine.” But the analogous sense for danger in the digital realm is far more limited. And this is where people can use a little help. We need someone to look over our shoulder and say “oh I really won’t provide that information to that site… it looks dodgy to me.” We need a noodge. A noodge is someone who pesters somebody to do something. And in this case, the noodge steps in to pester you into at least thinking before sharing that piece of data or browsing that site. This noodge could also inform the credential broker and data manager to create a pseudonymous email address and a one-time credit card number for use on a specific service. 

We have this to a limited degree today. Browsers provide indicators about the “safety” of a website. But imagine if we could scale up safe browsing data to all of our digital interactions including those brokered by apps and agents. A super powered noodge could help prevent people from falling for fraudsters and over-sharing data.

But these three things alone don’t get us where we need to go. So what’s missing

The Interface Layer

What is missing is a user interface layer that is more than a Teddy Ruxpin clone that only knows a few key phrases and actions accompanied by zero contextual awareness. What is needed is a meaningful conversational interface that is contextually aware. Generative AI and large language models (LLMs) are showing promise that they can power that layer. 

This interface layer must be multimodal allowing for audio, text, and tap inputs to be able to serve the majority of our global population. Good news is that LLMs are actually good at doing this.

Furthermore, this interface layer needs to run on form factors from mobile, to wearable, to implantable. There is strong evidence that the week after I present this, Apple will announce Siri running natively on its phone, without a need to phone home. And the push to have these models closer to the edge will only continue.

Ok so far we have an interface layer brokers by generative AI and LLMs, a credential broker, a data manager, and the noodge. But that too is still not enough to get to where we need to go. And here I have to thank Nat Sakimura for pointing this out to me.

Preference Resolver and Manager

What’s missing is privacy. More specifically, the combination of ways to express what data am I okay sharing for what purposes with clear statements from the service with which I interact on how they use that data. For example:

  • I am okay trading my email address for a one-time coupon.
  • I am okay sharing my blood pressure monitor data with my general practitioner.
  • I am okay presenting my student ID card to get into the movies cheaper. And so on.

Meanwhile services:

  • Can use your email address to notify you about new products and services
  • Will share your data with third party service providers
  • Will sell your data to data brokers

We need both in order for us, and the counselors, to make informed decisions. Thus we need a way to resolve what service will do with our data – these are privacy notices-  and we need a way of managing what we are okay with sharing under what conditions. Today, we do not have either in a consistent manner. We lack standards here. This is frankly a big gap.

Participatory Surveillance Manager

And now the last ingredient. This is the scariest one and the hardest one and, in some respects, the more important one.

In order for counselors to act on our behalf and truly understand what we want to accomplish, they need to observe what we typically do with our devices. Which apps do we use? Which sites do we visit? Which pieces of data do we share and where do we share? Which credentials do we use for which services and sites?

And in order to do this, something has to surveil us and we would need to be okay with it. Let’s let that sit there for a minute. In order for something to truly act on our behalf, it needs to study us, learn from us, and we need to be okay with that.

And we are okay with this concept in our lives today… don’t believe me? Your partner and your friends are examples of that kind of something. As is your executive assistant, if you are lucky enough to have one. You trust them. They know what you like and dislike. They can even notice when you are behaving a bit off from normal. That is participatory surveillance and counselors need it.

Consider that alternative – non-participatory surveillance. We have that today in the digital realm and we have a lot of it. Nation state surveillance. Employer surveillance. None of which we get an opt-in or opt-out for. 

Counselors to be truly useful need a very robust Preference Resolver & Manager coupled with Participatory Surveillance Manager. We have semblances of this today. Siri and Google Assistant are notionally bound to only use specific kinds of data from specific apps on our devices. This concept could be and would need to be scaled up.

Counselors in Summary

In order for counselors, these smart AI-powered agents, to be of use, they will need to rely heavily on things you and I work on everyday. Starting with the Credential Broker and Data Manager – these components give the agent its means of acting on our behalf. The Noodge adds a layer of protection from the less savory parts of the internet as well as injects pseudonymity along the way. The Preference Resolver and Manager helps make my data use wishes known and helps ensure the service provider the counselor interacts with are ones that use my day aligned to my wishes. All of these exist to varying degrees today. 

Add a multi-modal interface layer and the counselors gain a voice so to speak – a means of interacting with the individual, on every device from mobile to wearable to implantable.

Adding to all this a participatory surveillance capability and the counselor learns what our regular interactions look like and mimic them. It is what grants counselors agents – it’s the thing that makes these smart agents, actually smart. It’s a bit scary but it holds an enormous amount of potential.

Applied Identity Management in Summary

Counselors are an exercise in applied identity management. The exercise takes foundational building blocks consisting of things you and I work on every day and combines them to form something that would help people actually achieve real outcomes… far beyond logging into something. (Remember… [do the audience thing again] life is more than just logging into stuff.) This moves far beyond user account management into using a digital identity to get something done in both real and online worlds.

This exercise of applying identity management is not strictly one of customer identity and access management. It can be performed in workforce use cases. It can be performed in citizen-centric use cases. It can be performed in humanitarian and unbanked use cases. 

And it is incumbent upon us, those who really understand the building blocks, to help our organizations, be they private or public, think beyond user account management to identity management. It is incumbent upon us to apply identity management.

Thank you.