The internet was born without an identity layer.

As we’ve previously written, many have attempted to build one, but doing so is complicated. Microsoft’s Head of Identity Kim Cameron – who helped build Microsoft Passport in 1999 –  explains why in his Laws of Identity

Why is it so hard to create an identity layer for the Internet? Mainly because there is little agreement on what it should be and how it should be run. This lack of agreement arises because digital identity is related to context, and the Internet, while being a single technical framework, is experienced through a thousand kinds of content in at least as many different contexts, all of which flourish on top of that underlying framework. The players involved in any one of these contexts want to control digital identity as it impacts them, in many cases wanting to prevent spillover from their context to any other.

There’s a tension here today – while individual internet players want to prevent spillover of context, individual users want context spillover in ways it serves them. Today, each of us act as a human “clipboard” using our labor to translate context from one application to another when the task could clearly be offloaded to AI.

Applications refuse to help for the sake of our privacy.

We think identity gains particular importance in the age of AI.  If we’re to have Personal AI, it’ll be based on who we are – our identity.  

The representation we use will be of critical importance because AI operates on context – if we choose a bad representation any Personal AI could feel shallow and unhelpful. So much of our digital identity stems from a Foucauldian subjectivity, where we leave it to major platforms to set the context and frameworks to construct our digital representations – boxing us into how others have observed us in contexts foreign-controlled.

Every internet player also tactically has a different view of our digital identity – so any identity we build must interop to the systems, representations, and interfaces preferred by any player.

We also see efforts in Web3 to construct identity as a set of verifiable, possibly immutable attributes. For Personal AI applications, we see such efforts as misguided and even potentially dangerous.

We previously wrote that Personal AI is an identity. In modern internet platforms identity is upstream of most applications to make application state available across devices. This is how Apple makes your file system available across devices.

In this blog we’ll defend the opposite direction: Digital identity is personal AI.

‍

Deep Real Identity

Prevailing narratives of identity center around namespaces and verifiable attributes.

• It’s composed of self-sovereign verifiable immutable attributes
• It should be of a namespace we control. We are “Crosshatch” not “CH” or “Cross Hatch”.
• Large platforms have assigned us attributes – our digital identity is “the sum of all digitally available information about that person”

These narratives are tactically reasonable but lack philosophical depth to a point of being misguided.

Doc Searls suggests identity as Deep Reals, something that

is fully human, and can elicit any level of emotional response, as real humans tend to do.

actually given by self-sovereign “standardized ways of letting others know required facts about us” e.g.,

"I’m over 18,”
“I’m a citizen of this country"
“I have my own car insurance”
“I live in this state"
“I’m a member of this club.”

We realize technology is a game of tradeoffs, but “standardized required facts” is clearly a far cry from anything “fully human.”

‍

Identity is emergent

We take on a firm post-structuralist view of identity: Each of us is contingent, fluid and emergent.

“Identities are only simulated, produced as an optical ‘effect’ by a deeper play, that of difference and repetition.” Deleuze, 1968

We’re compelled by modern philosophy here, but we think this idea is pretty natural. On a run, Coach Angela Davis declares

“You’re either growing or you’re slowing down.

The old you might not have had it in you to handle this moment. But the new you? Is embracing the challenge. The new you is leaning into the challenge. Now is another opportunity that we will embrace. You know what it takes to win. You know how to get there. You’re starting to see how every big win sets you up for something bigger.

Spending time, energy or compute verifying an attribute of the old you while the new you is constantly emerging feels like an odd use of time and an affront to our capacity to become.

‍

Almost nothing essential

Modern theories of reference are no more friendly to the idea of people as a composition of attributes.

If you’re going to assign an attribute to a person, it should be necessary i.e., arising in all possible worlds that it exists. It should be essential i.e., without this property the object would not exist.  

Gold has atomic number 79. If not atomic number 79 then not gold.

Try to find essential properties of other objects or people and you’ll quickly find yourself frustrated. “Identity company from Miami” does not necessarily define Crosshatch.

‍

That’s subjective!

We’d go further and note that example facts that might be assigned are likely subjective a la Foucault i.e., arising from power structures that impose foreign-defined attributes onto us.  If identity is to be self-sovereign, the sovereignty ought to extend to the frameworks themselves that assign any attributes we might accept as composing of our identity.

The stakes of namespaces appear to pale in comparison to the stakes of any candidate Deep Real identity. Sure it’s sort of annoying that at CVS I’m identified by my old phone number even though I have a new phone number.  That Doc Searls goes by ‘Doc’ and not ‘David’, even though he was named by his parents as David. Superman is Clark Kent. It doesn’t matter that Lois doesn’t know it yet.

We reject Foucaldian subjectivity induced by modern tech platforms – we rather actively construct our own. Endeavoring to own (or verify!) platform observations of us reifies their platform construct and the behaviors or dynamics they induce. We can never truly own the data about us if we can’t own the data generating processes themselves.

Attempting otherwise feels like Stockholm Syndrome.

‍

Cruising

Privacy in public is an antiquated construct belonging to the unremarkable and culturally acceptable – we were not afforded any true cocoon of anonymity. Distressing stereotypes persist on the basis of how people look or act. Discrimination is not solved by staying out of sight.

‍

Toward our own identity

These are not merely academic distinctions. To the extent our world is increasingly mediated via Personal AI, the context that powers this mediation should be truthful to us. This mediation already risks flattening human beings to a bland conformity. A pathway to protecting ourselves is one where the context AI has is truly necessary, not just idiosyncratic or of verified Foucauldian subjective realization.

We are not just the sum of all data collected about us, we are contingent images of it.

Accepting deep real identity as a collection of verifiable attributes assigned to us by large tech platforms feels both inhumane and not real at all. We are clearly not a collection of “required facts.”

But AI poses a natural path to construct sovereign contingent images. A kaleidoscope of emergent becomings.

• Any collected context
• from any framework that collects data about us
• Projected onto any other surface
• For purposes desired by us

If there is a self-sovereign digital identity, the best candidate today is Personal AI. Personal AI also solves the troubles identified by Cameron, as AI enables dual control of digital representation, enabling individuals to selectively share AI projections of themselves in whatever context they choose.

If an application requires a different representation – data interoperability is just a prompt away. This form of Deep Real identity is delightfully tractable, as it’s just

• any AI
• Any context
• Any prompt
• Any surface

operated under dual governance of application and user.

Widely cited 2016 posing of self-sovereign identity in Coindesk takes its first principle of self-sovereign identity as Existence, namely that identity is based on a personal core that resists digitization. It seems like a DeepReal identity as by Personal AI need not contradict this, but operate fully contingent on it.

Such a fluid digital identity need not deny applications where verified credentials are required. It could just rather be that digital personal identity operates on a plane separate from verified credentials, which could have altogether different requirements.

We’re excited to release our progressively self-sovereign digital identity this summer. One that's emerging and fluid.

And resists verification.