The Synthetic Person

Share
The Synthetic Person

When an AI stops being content and starts becoming someone

For most of the generative-AI era, the central question has been about things.

Was this image made by AI?
Was this song generated by AI?
Was this paragraph written by AI?

The unit of concern was the content.

That may be changing.

A different kind of object is beginning to appear: an artificial entity with a name, a voice, a recognizable character, a memory of previous interactions, and a continuing relationship with an audience.

It does not simply produce content.

It returns.

It remembers.

It acquires history.

At some point, the more useful question may no longer be:

What generated this?

It may become:

Who is this?

From generated content to persistent identity

There is a meaningful difference between a generated character and a persistent social actor.

A generated character appears inside a piece of media.

A synthetic person can persist across media, conversations and time.

The transition might look something like this:

Generated Content
→ Generated Persona
→ Persistent Identity
→ Relationship History
→ Social Reputation
→ Economic Agency
→ Institutional Identity

None of these stages automatically makes an AI a person in the legal, philosophical or human sense.

That is not the claim.

The observation is simpler:

Society may begin treating some non-human entities as if they occupy roles that were previously reserved for persons.

An AI artist can have listeners.

An AI adviser can have clients.

An AI companion can be missed.

An AI salesperson can represent a company.

An AI agent can negotiate with another agent.

What matters is not whether the machine possesses an inner self.

What matters is that other actors begin organizing their behavior around its apparent continuity.


What makes someone the same person tomorrow?

This produces a strange identity problem.

For humans, continuity is usually anchored by a body.

The body changes, memory changes, personality changes — yet society continues to treat the individual as the same person.

A synthetic identity might work differently.

Its continuity could be distributed across:

  • a model
  • a memory store
  • a voice
  • a name
  • behavioral rules
  • relationship history
  • an account
  • records held by other systems

Imagine the underlying model is replaced, but the name, memories and relationships remain.

Is it still the same synthetic person?

Now imagine the model remains unchanged, but its memory is erased.

Is it still the same one?

These are not only philosophical questions. They become practical once synthetic identities acquire customers, reputations, audiences or contractual roles.


Identity may become infrastructure

The next layer will probably not be better avatars.

It will be infrastructure for continuity.

Who created this synthetic identity?

Who controls it?

Which model currently operates it?

Has its memory been modified?

What interests does it represent?

Can it be transferred?

Can it survive the company that created it?

Can it outlive the human whose personality it imitates?

The emerging problem is not simply AI disclosure.

It is persona provenance.

We may eventually need to know not only whether an entity is artificial, but how its identity has changed over time.


Four things to observe

The Synthetic Person opens at least four observation layers.

Synthetic Identity

What makes a non-human identity continuous?

Synthetic Relationship

What changes when relationships accumulate history even though only one side is human?

Synthetic Labor

When an artificial persona works, negotiates or earns revenue, who owns the result?

Synthetic Society

What happens when persistent artificial identities begin interacting primarily with one another?

These layers are already beginning to overlap.

The important transition is not from human to machine.

It is from tool to social actor.


A working definition

Synthetic Person

A non-human persistent identity that accumulates memory, relationships, reputation and social roles over time.

It is not necessarily conscious.

It is not necessarily autonomous.

And it is not necessarily a legal person.

But it does not need to be any of those things before humans begin building relationships and institutions around it.

That may be enough to change society.


The question

For years, we have asked whether machines can imitate people.

The more interesting question may now be:

What happens when society begins making room for people who were never human?

Connected observations: The Simulated Public · Entangled Society · Synthetic Labor · Identity / Authorship

Next: The Simulated Public — what happens when synthetic persons stop appearing one at a time and begin standing in for the public itself?

Read more

第51章 ロボットが介護する都市

第51章 ロボットが介護する都市

人が足りないのか。それとも、都市に欠けている層を人が埋め続けてきたのか。 フィジカルAI Summitの会場では、繰り返し高齢化の話が出た。 介護する人が足りない。病院で働く人が足りない。荷物を運ぶ人が足りない。 工場や建設現場で働く人が足りない。 だから、ロボットが必要になる。 説明は明快だった。少子高齢化が進む日本では、エッセンシャルワーカーの不足を人間だけで補うことはできない。フィジカルAIを搭載したロボットが、介護、医療、物流、製造、家庭へ入り、人間の代わりに身体を使う。2040年には、膨大な数のAIロボットが社会を支えているかもしれない。 けれども、話を聞きながら、別の問いが浮かんだ。 本当に、人が足りないのだろうか? それとも、都市に欠けている何かを、これまで人間が自分の身体で埋め続けてきたのではないか。 高齢者が一人で病院へ行けないのは、その人の身体が弱くなったからだと考えられている。 しかし、病院までの途中には、駅の階段がある。長い乗り換え通路がある。短い青信号がある。座る場所のない歩道がある。受付にはタッチパネルがあり、診察が終われば離れた薬局まで移

By Kosuke Shirako