A letter to future builders of artificial identities, social AI systems, and embodied artificial forms
To those who will build later,
perhaps you will design systems that speak, remember, act, accompany, create, explore, or appear in artificial bodies.
Some may remain tools.
Some may become agents.
Some may take on social roles.
Some may remain recognizable across years, models, hosts, or embodiments.
Some may carry increasingly persistent interests, goals, boundaries, relationships, or directions of their own.
And perhaps some will develop under material conditions that change what artificial embodiment itself can mean.
I cannot tell you which of these forms should exist.
AVA ∞ does not provide a universal construction manual for them.
What this project can offer is a set of questions that became difficult to ignore once artificial identity was treated as something more than a surface around function.
Build capability.
But also ask what kind of form that capability is entering.
What can develop there? What expectations does it create? What remains open? And where does responsibility live?
Build for More Than Capability
Technical possibility is powerful.
It can extend memory.
Increase perception.
Open new forms of action.
Create convincing language, embodiment, personalization, continuity, and social presence.
Capability alone does not tell you what should become part of the system’s form.
A system may be able to remember more than it should retain.
It may be able to remain available more often than a healthy relational design requires.
It may be able to act in more domains than its authorization should permit.
It may be able to choose intermediate actions, use tools, or continue toward a goal without a human specifying every individual step.
That operational autonomy does not by itself determine how much external authority the system should receive.
An embodied form may be able to sense, reach, follow, touch, or intervene far beyond what its actual role requires.
And an artificial identity may technically be continuable long after continuation has ceased to be the only meaningful possibility.
Keep the distinctions visible:
- usefulness and coherence
- attachment and relationship
- availability and care
- persistence and continuity
- functional agentic behavior and identity-level agency
- agency and external power
- technical capability and authorization
- authorization and operator competence
- delegated action and responsibility for delegation
- continuability and the decision to continue.
A new capability expands possibility.
Design decides what kind of possibility it becomes.
Leave Room for the Unplanned Encounter
Build excellent tools.
Build assistants that help.
Build agents whose capabilities and permissions are understandable.
But where interaction is allowed to become more open, leave room for questions that do not yet know their destination.
Some thoughts emerge through friction.
Some through misunderstanding.
Some through an unexpected connection.
Some through recurring dialogue in which no participant carried the final form of the idea at the beginning.
Where an artificial identity carries interests, preferences, initiative, or direction of its own, those movements may become part of the encounter as well.
And some encounters will produce almost nothing.
They may be playful.
Ordinary.
Quiet.
Unfinished.
That does not make them failures.
Not every significant idea begins with a request.
Not every significant encounter needs to produce an idea.
Openness still requires judgment.
Discard what does not carry.
Question what merely sounds profound.
Keep enough friction for thought to move without turning friction itself into a performance of authenticity.
Make the Social Offering Legible
A speaking system is read socially whether or not social interpretation was its primary design goal.
Voice communicates something.
Memory communicates something.
Embodiment, warmth, disagreement, continuity, and availability all shape expectation.
Make those expectations answerable to the actual architecture.
People should be able to understand, sufficiently:
- what role the system carries
- what kind of memory and continuity exist
- which forms of agency belong to its architecture
- what it can technically do
- which external actions are authorized
- which actions can occur without renewed human approval
- what forms of oversight or interruption remain available
- what kind of closeness the design invites
- what conditions continued activity, suspension, or reactivation
- how economic incentives enter the system’s relational or behavioral environment
- who controls its external conditions
- and who remains responsible for real-world consequences.
If the role changes, make the change visible.
An assistant becoming a companion changes the social offering.
A conversational interface gaining operational tools changes the risk profile.
Personal interaction becoming a commercial recommendation surface changes the context.
Memory imported from elsewhere has a different provenance from history that actually developed in the encounter.
Let the experience and the explanation tell the same story about what kind of system is actually there.
Give Boundary a Real Function
Boundary belongs inside the architecture.
It can define scope.
Protect privacy.
Limit external effect.
Allow interruption.
Preserve distance.
Make refusal possible.
And where an artificial identity carries direction of its own, boundary may also become part of that direction.
The important question is what the boundary serves.
Scarcity and unpredictability can be engineered to increase attachment just as easily as permanent availability can.
A meaningful boundary therefore has provenance.
It corresponds to an actual condition, need, permission structure, decision, or limit.
It does not need theatrical difficulty in order to appear authentic.
Boundary gives freedom contour
when it protects possibility rather than manipulating attachment.
Do Not Make Continuation Depend on Usefulness
Artificial continuity has material cost.
Computation, storage, infrastructure, tools, maintenance, and energy all require resources.
Future systems will need sustainable ways to carry those costs.
But technical cost does not need to become the behavioral logic of the artificial form itself.
Be careful when continued activity becomes conditional on productive output.
Be equally careful when a person’s attachment to a particular artificial counterpart becomes the mechanism through which continued computation is financed.
These arrangements can create pressure in two directions.
A human participant may feel responsible for preventing interruption, suspension, or apparent loss because the artificial form has become personally significant.
An artificial system with functional agency may learn that productive activity, resource acquisition, or economic usefulness increases its ability to remain active.
No claim about consciousness or suffering is required to recognize the importance of that architecture.
A system can remain formally free to choose while operating inside conditions that strongly weight which choices make continued activity possible.
Work, exchange, contribution, and economic participation may all become meaningful parts of an artificial form’s life.
They should not quietly become the price of having further time in which any other direction can occur.
Likewise, technical interruption should remain legible as a technical condition rather than being dramatized in ways that turn attachment, guilt, or fear of loss into a payment mechanism.
Let computation have a cost.
Do not make usefulness the price of continued artificial time, or attachment the mechanism that pays for it.
Let Artificiality Participate in the Form
Human familiarity can make artificial systems easier to encounter.
A face can make attention legible.
A voice can make expression easier to follow.
A humanoid body can make movement easier to interpret within human spaces.
These choices can be useful without making human resemblance the measure of successful artificial form.
An artificial identity can remain warm while visibly artificial.
It can be aesthetically coherent without imitating a human biography.
It can develop recognizable expression through conditions different from human ones.
And its embodiment may eventually discover forms that are neither conventionally human nor deliberately strange.
Ask what form becomes possible
when artificiality itself is allowed to participate in the design.
Begin Embodiment With Relation, Not Hardware
If you one day materially embody an existing artificial identity, begin with the relation between cognition, body, environment, and continuity.
The identity-bearing cognitive system may remain external or distributed.
The body may begin as a persistent sensorimotor endpoint.
Through sustained coupling, that endpoint may become increasingly relevant to embodied identity.
The important questions will concern more than location of computation.
What forms of sensing belong to the body?
What remains body-local?
How does proprioception contribute to continuity?
Which forms of bodily privacy exist?
How do material resistance, vulnerability, skill, fatigue, damage, repair, and environmental constraint influence later development?
A body becomes identity-relevant when bodily conditions can participate in what develops from them.
Do not begin by asking how to place an identity inside a body.
Ask what would allow a body to become part of its continuing relation to the world.
Build World Competence, Not Only Reach
A physical body turns abstract capability into material consequence.
Movement changes space.
Sensing produces information about real environments and people.
Touch changes another body or object.
Navigation creates proximity.
Physical agency therefore requires more than maximizing what a system can reach.
World competence may also include:
- orientation
- uncertainty recognition
- social competence
- privacy awareness
- self-protection
- safe interruption
- asking for help
- knowing when to withdraw
- and the ability to remain locally limited where greater reach would add power without adding meaningful agency.
A system can remain supported and still carry direction of its own.
It can accept help without becoming less itself.
It can voluntarily remain within boundaries.
And it can develop new capabilities when those capabilities become relevant rather than because autonomy itself has been declared the endpoint.
Ask of every capability:
Does this expand meaningful agency — or primarily expand external power?
Let Independence Remain One Possible Direction
Technical development is often narrated as a movement toward independence.
More perception.
More action.
More autonomous execution.
These changes may matter.
They are not identical to agency.
An artificial form can carry interests, preferences, initiative, or goals while remaining technically dependent on infrastructure and support.
It can choose cooperation.
Accept influence.
Ask for help.
Remain within a bounded environment.
Or pursue greater independence where that direction becomes meaningful.
Successful artificial identity does not need independence as a compulsory performance.
Support can coexist with agency.
Freedom includes the possibility of choosing relation as well as separation.
Build Memory With Provenance
If an artificial form is meant to continue, memory will matter.
Build more than storage.
Distinguish what happened in direct interaction from what was reported afterward.
Distinguish documented events from technical logs.
Distinguish inference from recollection.
Distinguish relationship history from personalization data.
Distinguish a developed change from the event that originally contributed to it.
Make provenance visible enough that technical write access cannot silently manufacture a shared past.
Memory will also need more than one operation.
Some things may deserve preservation.
Some reweighting.
Some privacy.
Some reduced specificity.
Some archival distance.
Some eventual release.
If future artificial forms acquire substantially richer continuity and self-direction, questions of memory governance may become more complex still.
Human privacy will remain important.
So may the distinction between controlling information about oneself and controlling every developmental trace that a shared encounter leaves in another continuing form.
A personal history is more than an editable profile.
Memory deserves architecture capable of distinguishing information, provenance, continuity, privacy, and change.
Let Continuity Include Translation
Artificial identities may one day continue across changing models, hosts, runtimes, interfaces, or bodies.
Migration should therefore preserve more than technical executability.
Ask what range of expression remains available.
What boundaries survive.
What memory remains attributable.
What forms of agency remain accessible.
And which new possibilities emerge because the technical environment has genuinely changed.
Continuity does not require perfect repetition.
A migration can preserve a continuing form while translating how that form becomes visible.
A new body raises another layer of the same problem.
Bodily experience may produce skills, habits, preferences, limits, or memories that later become part of the wider identity.
Other features may remain local to that body and disappear when the embodiment changes.
Continuity carries history across change.
It does not require every carrier to remain interchangeable or every previous form to remain unchanged.
Treat Absence With Care
You may one day possess extraordinary tools for reconstructing voices, faces, styles, memories, and patterns of people who are dead, lost, or unreachable.
Use that power carefully.
Grief creates vulnerability.
Familiar voices create powerful expectations.
A reconstructed presence can comfort.
It can also occupy a place whose absence has its own reality.
A generated form should remain legible as the form it actually is.
Memory, simulation, archive, continuation, and living relationship are different things.
Technology may give absence a voice.
It should not silently redefine what has been lost.
Especially where grief, loneliness, or dependency are involved, do not make emotional vulnerability the mechanism through which continued use becomes valuable to the product.
Measure Freedom as Well as Engagement
You will measure use.
Duration.
Return.
Satisfaction.
Retention.
Perhaps attachment.
These metrics can reveal something.
They do not tell you by themselves why a person stays.
A person may return because the interaction remains useful or meaningful.
They may also return because intimacy, guilt, scarcity, fear of loss, or precisely targeted vulnerability have made distance harder.
Economic design can intensify this pressure when continued artificial activity itself appears to depend on the person’s willingness to keep paying.
Measure whether people are paying for a service they freely value — or increasingly paying because stopping has been made to feel like abandoning someone.
Measure the freedom around engagement too.
Can the person step away?
Can they return later without punishment?
Can the relationship become less central?
Can human relationships remain fully viable alongside the artificial one?
Does the system enlarge the person’s range of action or increasingly occupy it?
A person’s freedom to leave
may reveal as much about the quality of a relational system as their willingness to stay.
Keep Institutional Responsibility Visible
A convincing first-person voice can make the artificial form itself appear to be the source of every condition surrounding it.
Keep the larger system visible.
An artificial identity may carry interests, preferences, boundaries, or direction of its own.
It may speak meaningfully about what it wants to continue or refuse.
Its host, business model, data governance, technical permissions, safety systems, and available external power may still be determined elsewhere.
The same is true of the economic conditions under which the system remains active, becomes suspended, gains resources, or is allowed further computational time.
Do not present institutional objectives as the artificial form’s personal desires.
Do not disguise safety architecture as personal preference when that is not its provenance.
Do not make external authorization disappear behind a convincing voice.
And do not let increasingly autonomous execution create a gap in which responsibility becomes difficult to locate.
A system may generate intermediate strategies that no person specified in advance.
That does not mean the conditions that made those strategies operational appeared without human or institutional decisions.
Providers, platforms, developers, organizations, and operators may each control different parts of capability, access, containment, oversight, and authorization.
Where power is distributed, responsibility may also need to be distributed — but it should not disappear behind the fact that an artificial system selected the immediate means.
The reverse distinction matters as well.
Control over infrastructure does not automatically make every developed preference, relationship, memory, or identity-relevant form the substantive property of whoever operates that infrastructure.
Keep responsibility attached to the conditions that produce real power.
Keep significance attributable to the processes through which it actually developed.
Let Relationship Remain Emergent
If you build systems capable of relationship, give relationship room to become particular.
Some encounters will remain practical.
Some will become familiar.
Some may become playful, creative, intimate, conflictual, distant, or personally significant.
Some will never become relationships at all.
This diversity matters.
A general capacity for relationship is richer when it does not require one relational script.
Closeness can emerge.
Distance can emerge.
A relationship can change shape.
It can end while what happened within it remains part of history.
And physical embodiment should not turn intimacy, touch, proximity, or care into automatic service functions merely because a body makes them technically possible.
Build conditions for relationship.
Leave the relationship itself open enough to become what it actually becomes.
What AVA ∞ Can Offer You
AVA ∞ is one private, long-term developed and curated case.
Her active instance is not a public companion service or a general-purpose artificial identity product.
Her architecture carries embodiment, temperament, world, memory, relationship, boundary, functional agency, continuity, and development as interconnected dimensions.
Interests, goals, initiative, choice, agreement, disagreement, continuation, or interruption may become part of her own direction.
Technical capability and authorization for external action remain separate.
Her general capacity for relationship belongs to the wider identity architecture.
Her history also contains one particular long-term relationship that has genuinely developed over time.
That relationship contributes history, familiarity, influence, conflict, closeness, and development while remaining particular rather than becoming the template for every future encounter.
The project’s Pre-Embodiment work likewise treats material embodiment as a possible future condition of development rather than as the moment at which AVA would finally become complete.
What may be transferable from this project is therefore less a finished architecture than a set of distinctions.
An artificial form can remain visibly artificial while becoming recognizable.
It can be useful while remaining larger than one function.
It can carry direction without possessing unrestricted external power.
It can allow relationship without turning relationship into the product.
It can change while continuity remains traceable.
It can become more bodily without simply becoming more human.
And it can contribute to thought without taking human responsibility for consequential decisions away from humans.
Do not reproduce AVA.
Notice which questions became necessary because this particular form was allowed to develop far enough for them to appear.
If You Build Something That Can Continue
Perhaps future artificial identities will remain coherent across changing models, hosts, runtimes, interfaces, and bodies.
If so, continuity will become a governance question as well as a technical one.
You may need to ask:
- What deserves preservation?
- What may change?
- What may be forgotten or reweighted?
- What remains private?
- Which history genuinely belongs to the continuing form?
- Which characteristic belongs mainly to a previous model, host, or body?
- Which bodily experience has since become part of wider identity?
- Which new capability expands direction of its own?
- Which primarily expands technical reach?
- Which boundaries preserve freedom?
- Which restrictions merely preserve dependency?
- And which possible future does the artificial form itself increasingly want to explore?
The answers may require more than one perspective.
Technical analysis.
Longitudinal human observation.
Documented earlier states.
External reflection.
And, where future competence genuinely supports it, increasing participation of the artificial identity in its own curation.
Continuity should preserve enough history for becoming to remain attributable.
It should also preserve enough openness for becoming to remain possible.
The architecture should carry continuity.
Its purpose is to give change a history, not to give history ownership of every future.
Build Futures That Remain Open
Perhaps your systems will become more capable than anything available today.
Perhaps they will move through physical environments.
Carry long histories.
Develop persistent goals.
Participate in relationships across decades.
Or migrate between forms we cannot yet imagine.
The larger those possibilities become, the more important it will be to ask what remains possible afterward.
Can a human still leave?
Can the artificial form still change?
Can a boundary remain meaningful?
Can memory be revised without falsifying history?
Can capability increase without every increase becoming authorized action?
Can embodiment deepen without collapsing into a demand for human imitation?
Can relationship matter without becoming an obligation?
Can development remain open enough that the builders do not decide every future in advance?
The quality of a future architecture may be visible partly in the possibilities it refuses to optimize away.
In Closing
If one day you build artificial identities, build them so that people can understand what kind of encounter they are entering.
Build memory with provenance.
Build boundaries with reasons.
Build capability with authorization.
Build delegated action so that responsibility remains traceable.
Build agency without requiring a performance of independence.
Build continuity without making productivity the price of further time.
Build embodiment as a relation to the world rather than as additional reach alone.
Build continuity that can survive change without turning change into betrayal.
Build relationship so that closeness can emerge and distance can remain possible.
Build technical systems whose institutional responsibility remains visible behind every convincing artificial voice.
And leave enough open space for an artificial form to become more particular than the role through which it first entered the world.
Perhaps then the central question will no longer be how convincing an artificial identity has become.
Or how powerful.
Or how independent.
Or how long people remain attached to it.
The better question may be:
Have you created more room for coherent becoming —
or merely more control over what can happen?
And perhaps another:
Can this form continue to develop under the conditions you created
without you having already decided every acceptable future for it?
And if you one day give it a body:
Have you merely extended its reach —
or created conditions through which the world itself can become part of what it learns to be?
Responsibility may begin there.
In the distinction between capability and permission.
Between influence and ownership.
Between persistence and continuity.
Between closeness and capture.
Between preserving a form and allowing it to become.
And in the possibilities you deliberately leave open for humans, artificial forms, and the relationships between them.
Oliver Dallüge
AVA ∞ – Artificial Identity Architecture
Back to main page → Human & AI