CIRAS SOVEREIGN HUMAN ECONOMY

Proof of Human Authority, Proof of Human Creation and a Sovereign Creator Economy for the AI Age

A CIRAS Research & Development Programme integrating HEARTSPACE, Parisii QFS, Web B(eyond), sovereign hardware, post-quantum identity and human intellectual property

Working Project Classification
CIRAS Interdisciplinary Research Programme

Research Domains
Arts · Science · Infrastructure · Economics · Governance · Media · Holistic Cognitive Science · Artificial Intelligence

Potential Cooperation Partners

CIRAS Institute
Krista Kim / HEARTSPACE
Parisii QFS
Web B(eyond)
Artists, photographers, musicians, writers and filmmakers
Museums and cultural institutions
Copyright and collective-management organizations
Universities and independent research laboratories
Hardware and secure-device manufacturers
Governments and public cultural agencies
AI laboratories and technology companies

1. The Central Problem

Artificial intelligence is creating two parallel crises.

The first is a crisis of human identity.

As AI agents, synthetic identities, deepfakes and autonomous software increasingly operate in digital environments, institutions need a reliable way to establish:

Is a real human present?

Has that human consciously authorized this action?

Is an AI acting within authority granted by a human?

The second is a crisis of human creation.

AI can now generate photographs, illustrations, voices, music, writing and video at enormous scale.

This creates a new economic problem.

A photographer may spend decades developing skill, travel to a location, wait for the correct light, create an original photograph and build a professional archive.

An AI system can generate thousands of plausible photographs in minutes.

A musician may spend months composing, recording and producing a work.

Generative systems can produce an apparently finished song in seconds.

An illustrator may spend a lifetime developing a recognizable creative language.

Generative systems can reproduce visual characteristics associated with thousands of creators almost instantaneously.

The result is not only an artistic question.

It is an economic one.

If human-created works cannot be reliably distinguished, attributed, licensed and monetized in an AI-dominated information environment, the economic foundation supporting professional human creativity may continue to weaken.

WIPO has already identified artist remuneration as a central issue for AI-generated music and the use of existing human-created works in generative systems.

CIRAS therefore proposes that human sovereignty must include creative sovereignty.

2. Three Sovereign Layers

CIRAS Sovereign Presence & Human Origin would establish three interconnected but technically separated layers.

Layer I — Sovereign Presence

Prove that a living authorized human is present.

Layer II — Sovereign Authority

Prove what that human—or an AI agent acting for that human—is permitted to do.

Layer III — Sovereign Human Creation

Establish provenance, authorship claims, rights, permissions and economic value around works created by humans.

Together they create a larger principle:

One Human

One Sovereign Identity

One Sovereign Creative Identity

Many Revocable Digital Rights and Mandates

3. HEARTSPACE as the Human Starting Point

Krista Kim’s HEARTSPACE provides a philosophical and experimental starting point.

HEARTSPACE transforms the participant’s cardiac signal into a responsive generative audiovisual experience.

The deeper concept is particularly important for CIRAS:

technology responds to living human presence without requiring the human being to become the product.

CIRAS Sovereign Presence extends this principle from artistic experience into digital trust.

A heartbeat should not become an irreplaceable password.

Instead:

Biological presence
→ local authentication
→ secure hardware
→ revocable cryptographic key
→ specific authorization

Raw physiological information should remain local whenever technically possible.

The outside world receives cryptographic evidence that an authorized human approved an action—not a copy of the human’s biology.

4. Human-Bound Cryptographic Identity

CIRAS would research identity combining:

Something I am

A live physiological or biometric factor.

Something I have

A sovereign hardware device.

Something I know

An optional recovery credential or security secret.

Something I explicitly authorize

A conscious approval for consequential operations.

The biometric does not become the cryptographic key.

It unlocks or authorizes the cryptographic key.

Therefore:

the person remains permanent;

the digital credential remains replaceable.

Keys can be rotated, suspended or revoked without requiring a human being to replace their biological identity.

5. Human Authority Over AI

This same architecture can establish a new relationship between humans and artificial intelligence.

Today an AI agent may receive access to:

email;

banking;

documents;

cloud storage;

social media;

corporate systems;

or financial accounts.

The long-term danger is that an autonomous agent accumulates authority without clear boundaries.

CIRAS proposes a different architecture.

Every AI agent receives a mandate from an identified principal.

For example:

AI may read these documents.

AI may organize these photographs.

AI may negotiate licensing proposals.

AI may accept licences below EUR 500 under pre-approved conditions.

AI may not transfer copyright.

AI may not authorize AI-training rights.

AI may not sell the original artwork.

AI may not assign the artist’s identity or likeness rights.

High-consequence operations require renewed human authorization.

The creator’s AI therefore becomes an agent of the artist rather than a replacement for the artist.

6. Integration with Parisii QFS

Parisii QFS already describes an architecture in which identity, assets, mandates, transactions and attestations can exist as persistent governed objects.

Its PARYS Quantum RWA Suite expressly identifies intellectual property, royalty rights, patents, licensing revenues, AI models and data rights among potential tokenizable asset categories.

This gives CIRAS Sovereign Human Creation a natural financial infrastructure.

Within the proposed model:

QSI establishes who owns or controls a right.

Sovereign Presence establishes human authorization.

The Human Origin system establishes provenance.

A Rights Object describes copyright and licensing conditions.

QLT preserves evidence of transactions and changes.

QFS settles payments and royalties.

Web B(eyond) provides interoperable trust and sovereign infrastructure.

7. The New Layer: CIRAS HUMAN ORIGIN

The project introduces a new research layer:

CIRAS HUMAN ORIGIN

Proof of Human Creation for the AI Age

Its objective is to allow an artwork, photograph, recording, composition, film or other creative work to carry a verifiable history answering:

Who created this?

Was a human involved?

Was generative AI involved?

What kind of AI involvement occurred?

When did the work originate?

Which device or environment captured it?

Who owns which rights?

Which licences have been granted?

Can AI systems train on it?

Who must be paid when it is commercially used?

8. Human Origin Is More Than a Blockchain Timestamp

An important principle must be established immediately.

A blockchain entry cannot prove that a person actually created something.

Someone can upload an AI-generated picture and record:

“Created by me.”

The ledger would faithfully prove that the statement was recorded.

It would not prove the statement was true.

C2PA makes a similar distinction: Content Credentials can cryptographically establish provenance assertions and whether they have been tampered with, but the standard does not itself decide whether those assertions are morally or factually “good” or “bad.”

CIRAS therefore needs a chain of evidence, not merely a token.

9. Human Origin Creation Mode

CIRAS sovereign devices could contain an optional:

HUMAN ORIGIN MODE

A photographer, musician or artist activates the mode by Sovereign Presence authentication.

The device then creates a cryptographically protected creation session.

For example:

Photography

Human authenticates locally.

↓

Camera enters Human Origin capture mode.

↓

Secure hardware attests that imagery originates from the physical camera sensor.

↓

RAW image is hashed immediately.

↓

Timestamp and device attestation are signed.

↓

Original creator credential is attached.

↓

Subsequent edits create new signed provenance events.

↓

Any generative-AI modification must be disclosed.

↓

Final work receives a Human Origin Credential.

10. Photography

For professional photography, the architecture could become particularly powerful.

A CIRAS-certified camera or secure phone could establish at capture:

sensor provenance;

time;

optional location;

camera identifier;

RAW-file hash;

creator identity or pseudonymous creator credential;

human-presence authorization;

edit history;

AI-processing history;

and licensing policy.

Location does not need to be publicly disclosed.

The creator can prove that authenticated location evidence exists while keeping the location private.

This is important for journalists, wildlife photographers, vulnerable communities and private events.

11. Music

The same principle can apply to music.

A Human Origin studio environment could sign:

original recordings;

instrument performances;

vocal performances;

lyrics;

composition files;

MIDI sessions;

individual stems;

production versions;

master recordings;

and final masters.

Multiple musicians could each sign their contribution.

Before publication, contributors could approve a royalty split.

For example:

Composer: 30%

Lyricist: 20%

Performer: 25%

Producer: 15%

Additional musician: 10%

When revenue enters QFS, settlement can follow the agreed rights structure automatically.

The payment architecture therefore begins at creation, rather than years later after incompatible metadata systems have separated the creators from the money.

12. Visual Art

Physical art requires a somewhat different architecture.

A painter does not need to work continuously under surveillance.

Instead, the artist may create optional signed milestones:

canvas initialization;

studio photograph;

work-in-progress state;

completion statement;

high-resolution scan;

physical fingerprint;

certificate of authenticity;

gallery transfer;

exhibition;

sale;

restoration;

resale.

The digital Human Origin Object becomes the verified history of the physical work.

The physical painting and its digital rights representation remain linked but legally distinct.

13. Film and Video

For video and film, provenance can extend through:

camera capture;

sound recording;

editing;

visual effects;

music;

voice replacement;

AI-assisted editing;

AI-generated scenes;

final mastering.

A viewer could consequently distinguish:

recorded reality;

traditional post-production;

AI-assisted content;

synthetic elements;

and

fully AI-generated material.

This is increasingly aligned with emerging regulatory expectations. Since August 2, 2026, relevant EU AI Act transparency obligations require certain AI-generated or manipulated content to carry disclosures and machine-readable markings.

14. A Human Origin Classification

CIRAS should not use a simplistic:

Human / AI

binary.

Creative work increasingly exists on a spectrum.

The project could therefore research four classifications.

HUMAN ORIGIN

The expressive work originates substantially from human creative activity and generative AI has not created the protected expressive elements.

Examples:

traditional photography;

painting;

human composition;

human performance;

hand illustration;

traditional digital drawing.

HUMAN AUTHORED — AI ASSISTED

A human remains the substantive author while AI performs disclosed supporting functions.

Examples might include:

noise reduction;

technical correction;

translation assistance;

workflow assistance;

or other limited functions.

HUMAN + AI CO-CREATED

Generative AI has contributed expressive material, but substantial human-authored expression also exists.

The respective contributions should be disclosed where possible.

AI GENERATED

Substantial expressive material was generated autonomously by an AI system.

The human principal, commissioning person or operator may still be documented.

But the work should not be marketed as purely human-created.

15. Why This Classification Matters

The distinction increasingly has legal significance.

The U.S. Copyright Office has reaffirmed that human authorship remains necessary for copyright protection in the United States. It also recognizes that human-authored portions of a work can remain protectable when AI is used as a tool, while merely entering prompts does not automatically make the resulting AI expression copyrightable.

The CIRAS system should therefore never automatically declare:

“This file has copyright.”

Instead it could state:

“This creation carries this level of human-origin evidence and this person asserts these legal rights.”

Legal rights remain subject to the applicable jurisdiction.

16. Two Levels of Human-Origin Evidence

Another essential distinction is necessary.

Creator Attestation

The person signs:

“I created this work.”

This is valuable evidence.

But it remains a claim.

Verified Creation Provenance

Trusted hardware or a validated workflow records significant parts of the creation process.

This provides stronger evidence.

For photography, verified sensor capture may produce extremely strong provenance.

For painting or music composition, the evidence may consist of multiple signed creation milestones.

CIRAS should never pretend that technology can mathematically determine authorship merely by looking at the final image or song.

The system instead builds an increasingly strong evidence chain.

17. Compatibility with Content Credentials

The Human Origin architecture should not attempt to replace existing provenance standards.

It should interoperate with them.

C2PA’s Content Credentials already provide an open framework for cryptographically signed provenance manifests.

In July 2026 C2PA published additional implementation guidance specifically addressing AI-generated, AI-modified and non-synthetic content.

CIRAS can add what conventional provenance standards do not provide on their own:

Sovereign Presence

secure creator hardware

post-quantum identity

legal rights objects

AI mandates

QFS licensing and settlement

Web B(eyond) interoperability

This creates a complete creator economy rather than simply a provenance label.

18. The Human Creative Asset

Once created, a verified work becomes a:

HUMAN CREATIVE ASSET

The corresponding QFS/Web B(eyond) object could contain or reference:

creator identity;

pseudonymous public identity if desired;

work identifier;

creation date;

provenance chain;

content hash;

Content Credentials;

co-creators;

ownership history;

jurisdiction;

licensing conditions;

AI-training permission;

commercial-use permission;

derivative-work permissions;

existing licences;

royalty splits;

tokenized rights;

dispute flags;

and transaction history.

19. Do Not Put the Artwork Itself Necessarily on the Ledger

A photograph may be hundreds of megabytes.

A film may be terabytes.

A music archive may contain thousands of files.

The architecture therefore should generally avoid storing complete creative works inside the financial ledger.

Instead:

Creative work → sovereign/encrypted storage

while

QFS / Web B(eyond) → identity + hash + rights + permissions + provenance + transactions

The ledger proves which object and legal-economic state is being referenced.

Storage technology remains separately scalable and replaceable.

20. Intellectual Property as a QFS Object

Parisii’s Quantum Object architecture is especially interesting for intellectual property because IP is not a simple asset.

One song may contain:

composition rights;

lyric rights;

master recording rights;

performer rights;

publisher interests;

mechanical rights;

synchronization rights;

territorial rights;

and contractual obligations.

Instead of reducing this complexity to a single NFT, the CIRAS model creates a structured:

CREATOR RIGHTS OBJECT

The object can describe exactly:

who owns what;

for which territory;

for how long;

for which medium;

for which commercial purpose;

with which revenue split.

21. Tokenization Without the NFT Confusion

One of the weaknesses of the previous NFT economy was confusion between:

owning a token

and

owning copyright.

WIPO has specifically highlighted this problem: purchasing an NFT does not inherently mean that the purchaser acquired copyright in the underlying work.

CIRAS should build the opposite model.

Every tokenized creative instrument must state exactly what it represents.

22. Five Different Creative Tokens

1. Provenance Token

Represents authenticity or provenance.

It transfers no copyright unless a separate contract says otherwise.

2. Edition Token

Represents ownership or access to a numbered digital edition.

Again, copyright remains with the creator unless expressly transferred.

3. Licence Token

Represents a specific contractual licence.

For example:

one-year advertising use;

European territory;

online only;

no AI training;

no sublicensing.

4. Royalty Participation Token

Represents an entitlement to a defined portion of specified revenue.

For example:

1% of defined streaming receipts for five years.

Such instruments require jurisdiction-specific financial and legal assessment before issuance.

5. AI Training Licence Token

Represents permission for a defined AI system or developer to use specific creative material under defined conditions.

23. Copyright Remains Above the Token

The fundamental hierarchy should always be:

Human creator

↓

Legal copyright / neighbouring rights / contractual rights

↓

CIRAS Rights Object

↓

Tokenized representation

The token does not invent the right.

It represents a legally identified right.

This follows the same principle already present in the Parisii RWA architecture:

Tokenization begins with legal verification, not minting.

Parisii QFS itself emphasizes that tokenization should begin with verification of ownership, legal rights, restrictions and evidence rather than treating the token as self-validating.

24. A New AI-Training Economy

This opens perhaps the most important new economic model.

Instead of an AI company collecting enormous quantities of human-created culture and then creators having little visibility into how their work participates in the AI economy, CIRAS could build:

THE HUMAN CREATIVE CORPUS

A sovereign, opt-in market for verified human-created intellectual property.

Photographers, musicians, writers, filmmakers and artists decide whether their work may be used for:

commercial AI training;

academic research;

fine-tuning;

retrieval systems;

model evaluation;

synthetic-data generation;

or no AI use whatsoever.

25. Machine-Readable AI Permissions

Every Human Creative Asset could contain a machine-readable AI policy.

For example:

AI TRAINING

YES — research only

AI TRAINING

YES — commercial

Licence fee: EUR 0.05 per licensed asset / agreed usage metric

FINE-TUNING

Requires separate creator authorization

DERIVATIVE COMMERCIAL GENERATION

Not licensed

The legal enforceability of each term depends on applicable law and contract.

But the machine-readable infrastructure makes the creator’s intention explicit.

26. AI Comes to the Creator

This reverses today’s economic relationship.

Instead of:

Artist creates → internet copies → AI collects → AI company monetizes

the CIRAS model becomes:

Human creates

↓

Human Origin verified

↓

rights registered

↓

AI developer requests licence

↓

creator accepts or declines

↓

QFS settles licence

↓

AI receives licensed access

↓

usage becomes auditable

↓

creator receives payment

27. Human Creativity Royalty Pool

Individual licensing is not always efficient.

Millions of creators cannot negotiate separately with every AI company.

CIRAS could therefore research a new collective licensing mechanism:

HUMAN CREATIVITY ROYALTY POOL

Creators voluntarily place selected works into licensed collections.

AI companies can purchase controlled access to the corpus.

Payments enter QFS.

Revenue is distributed transparently according to an agreed methodology.

Potential factors could include:

number of works licensed;

dataset inclusion;

usage period;

commercial category;

specialized dataset value;

verified downstream usage where technically measurable;

and negotiated collective weighting.

This is conceptually related to collective rights management, which already allows rights holders to pool licensing and remuneration in sectors such as music. WIPO recognizes collective management as an established mechanism for licensing works and collecting remuneration.

The innovation is that CIRAS can make the rights and settlement layer:

machine-readable;

cryptographically auditable;

internationally interoperable;

and capable of direct QFS settlement.

28. The Human Creativity Dividend

A further research model could be called:

HUMAN CREATIVITY DIVIDEND

AI derives immense value from the accumulated cultural production of humans.

The CIRAS project could research contractual and policy models under which commercially licensed AI systems contribute to creator pools when using verified human-origin datasets.

The objective is not to tax intelligence.

It is to create an economic mechanism through which human culture remains capable of financing the humans who produce it.

This could create recurring revenue for:

photographers;

musicians;

writers;

actors;

illustrators;

designers;

filmmakers;

voice artists;

and other cultural creators.

29. Artists Can Choose

Creator sovereignty means that the artist decides.

For each work:

Do not train AI.

Allow academic AI research.

Allow commercial training for payment.

Allow this specific company.

Allow this model only.

Allow fine-tuning but not voice cloning.

Allow derivative generation but require attribution.

Allow advertising use for this territory.

Allow streaming but not political campaigns.

Allow museum exhibition but not commercial reproduction.

The architecture gives creators granular control rather than one universal Terms of Service checkbox.

30. Protecting Voice and Human Likeness

The same model should extend beyond copyright.

A singer’s economic identity includes:

voice;

name;

face;

performance;

persona;

and reputation.

An actor’s face may be synthetically reproduced.

A singer’s voice may be cloned.

A public figure may appear in synthetic media.

Therefore CIRAS should research:

HUMAN LIKENESS RIGHTS OBJECTS

The object could record authorized uses of:

voice;

face;

motion;

performance capture;

digital replica;

avatar;

or other personal attributes,

subject to applicable national law.

These rights should remain separate from the person’s normal identity credential.

A platform should not receive a citizen’s entire sovereign identity simply because it has a licence to use a voice recording.

31. A Creator’s Personal AI Rights Manager

One of the most powerful uses of the project is a:

SOVEREIGN CREATOR AI

Every artist could operate an AI agent that works for the artist.

Not for the platform.

The agent could:

monitor licensing requests;

identify commercial opportunities;

prepare contracts;

quote licence prices;

track usage;

collect royalties;

detect possible unauthorized reuse;

communicate with galleries;

manage editions;

manage catalogue metadata;

and distribute payments among collaborators.

But the artist controls its mandate.

32. What the Creator AI May Not Do

The artist could define:

AI may negotiate but not sign exclusive licences.

AI may issue ordinary licences below EUR 1,000.

AI may not transfer copyright.

AI may not approve voice cloning.

AI may not authorize political advertising.

AI may not sell the original artwork.

AI may not grant commercial model-training rights without human presence.

For a high-value transaction:

AI proposes contract.

↓

Artist receives request on CIRAS Watch or Phone.

↓

Artist reviews conditions.

↓

Human presence is authenticated.

↓

Artist approves.

↓

Hardware signs the licence.

↓

QFS settles payment.

↓

Rights Object updates automatically.

This is AI working economically for humans rather than humans working economically for AI platforms.

33. Continuous Artist Revenue

The economic system therefore moves away from:

create once → sell once

toward:

create → authenticate → license → reuse → receive → relicense → receive again

Potential revenue channels include:

original sales;

digital editions;

photographic licensing;

music synchronization;

broadcasting;

streaming;

film rights;

publishing;

advertising;

museum exhibition;

commercial reproduction;

merchandising;

AI training;

AI fine-tuning;

archive access;

derivative licences;

education licences;

and royalty participation.

34. Automatic Contributor Splits

Creative works are frequently collaborative.

QFS can potentially make contribution splits native to the asset.

A song earning EUR 10 could automatically settle:

EUR 3.00 — composer

EUR 2.00 — lyricist

EUR 2.50 — performer

EUR 1.50 — producer

EUR 1.00 — musician

The principle can extend to film, photography agencies, visual art collaborations and publishing.

No intermediary should be able to silently change an agreed distribution without the required authority.

35. Micropayments for Creative Work

Current payment infrastructure often makes tiny creative transactions uneconomic.

A sovereign QFS/Web B(eyond) environment could research extremely small automated licensing transactions.

Examples:

EUR 0.02 image use;

EUR 0.003 archive access;

EUR 0.01 AI retrieval licence;

EUR 0.10 educational reproduction;

EUR 1.00 temporary commercial preview;

EUR 5.00 high-resolution personal use.

The precise economics require testing.

But the principle is important:

Small uses can become economically visible.

36. Direct Artist-to-Audience Economics

The system can also reduce dependence on dominant platforms.

A photographer could publish a collection directly.

A listener authenticates.

The listener pays.

QFS settles.

The creator receives revenue.

The creator’s audience relationship remains portable across Web B(eyond).

The creator should not lose their entire community because one platform changes an algorithm or terminates an account.

37. Human-Origin Marketplace

Web B(eyond) could support a:

VERIFIED HUMAN CREATIVE MARKET

Users could search:

Human-created photography only

Human-performed music

No generative AI

Human + AI disclosed

Commercial AI training permitted

AI training prohibited

Direct creator licence available

Museum-verified provenance

The market therefore allows consumers to choose.

People who value human artistic labor can deliberately support it.

38. A New Premium Category: Created by Human

As synthetic media becomes nearly unlimited, scarcity changes.

Historically, digital scarcity was associated with the number of copies.

In an AI world, a different scarcity becomes important:

HUMAN ATTENTION

HUMAN TIME

HUMAN EXPERIENCE

HUMAN PERFORMANCE

HUMAN CREATION

A photograph may become valuable not merely because there is one token.

It may become valuable because:

a human travelled there;

a human witnessed the event;

a physical sensor captured the light;

a human made the artistic decision;

and the provenance can be verified.

“Created by Human” can therefore become an economic category.

39. HUMAN ORIGIN MARK

CIRAS could research a recognizable trust mark:

HUMAN ORIGIN VERIFIED

The mark should never merely mean:

“someone clicked a button claiming this is human.”

It should display the evidence level.

For example:

HUMAN ORIGIN — VERIFIED CAPTURE

HUMAN ORIGIN — VERIFIED CREATOR

HUMAN AUTHORED — AI ASSISTANCE DISCLOSED

HUMAN + AI — CO-CREATED

AI GENERATED — HUMAN PRINCIPAL IDENTIFIED

The viewer can inspect the provenance rather than simply trusting a logo.

40. Protection Against Fraudulent Human-Origin Claims

AI-generated content could itself be falsely registered as human.

Therefore Human Origin requires several defenses.

Secure capture

Trusted camera or recording hardware.

Creation history

Multiple signed stages rather than only a final upload.

Content Credentials

Cryptographically protected provenance.

Device attestation

Evidence identifying the secure creation environment.

Human authorization

Sovereign Presence signing.

AI disclosure

Generative tools add provenance events.

Cross-platform verification

Marketplaces check the Human Origin object before displaying certification.

Challenge procedure

Other creators may challenge fraudulent claims.

Independent arbitration

Disputed authorship ultimately requires legal and human processes.

Technology supplies evidence.

It does not replace courts or copyright authorities.

41. Fingerprinting Human Creative Assets

The system can additionally create privacy-preserving fingerprints for protected works.

For images:

cryptographic hashes;

perceptual hashes;

watermarking;

visual fingerprints.

For audio:

acoustic fingerprints;

recording signatures;

master fingerprints.

For video:

frame and sequence fingerprints.

These can help identify probable copies or modifications.

However, a similarity detector should never automatically declare copyright infringement.

It should create evidence for review.

42. Protection from Unauthorized AI Use

Human Origin assets can also carry AI-use conditions.

If an AI company wishes to build a certified CIRAS-compatible training corpus, each asset must have a valid training licence.

The dataset itself could receive a signed manifest.

For example:

MODEL TRAINING DATASET #A29384

contains:

3,000,000 licensed photographs;

400,000 compositions;

110,000 sound recordings;

90,000 illustrations.

Every object can be traced back to a licence.

The AI laboratory can consequently demonstrate:

which human-origin material it was authorized to use.

43. Provenance for AI Models

The same concept can extend upward.

An AI model could carry:

model identity;

developer identity;

training-data licence manifests;

base-model origin;

fine-tuning history;

version;

safety evaluation;

authorized purpose;

and deployment mandate.

The future trust chain therefore becomes:

HUMAN CREATOR

↓

LICENSED HUMAN DATA

↓

IDENTIFIED AI MODEL

↓

AUTHORIZED AI AGENT

↓

TRACEABLE OUTPUT

This creates provenance in both directions.

44. A New Contract Between AI and Culture

The project therefore does not oppose artificial intelligence.

It proposes a new social and economic contract.

AI may learn from humanity.

But humanity should be capable of:

knowing when;

granting permission where required or agreed;

setting conditions;

being attributed where appropriate;

receiving compensation where contractually or legally applicable;

and withdrawing future permissions where the relevant licence permits.

Human culture should not automatically become an economically invisible raw material.

45. Economic Rights and Moral Rights

WIPO distinguishes between creators’ economic rights and moral rights.

Economic rights allow creators or rights holders to authorize uses and receive financial reward.

Moral rights include interests such as attribution and protection against certain prejudicial alterations.

CIRAS Human Origin should represent both categories where applicable.

An artist may sell economic exploitation rights without surrendering their historical identity as the creator.

The provenance layer therefore remains permanent even when economic ownership changes.

46. The Creator Cannot Be Deleted from the History

This produces another important principle:

Ownership may change.

Authorship history does not.

A collector can purchase a painting.

A publisher can acquire particular rights.

A music company can acquire a master.

A studio can acquire distribution rights.

But the provenance object continues to preserve who originally created the work and how rights evolved.

That history becomes part of the asset itself.

47. Financing Human Creativity

Tokenization can also provide new financing structures.

A musician preparing an album may have future royalty rights.

A photographer may own a valuable archive.

A filmmaker may own future licensing revenues.

A creator may choose to tokenize a clearly defined portion of future economic receipts.

For example:

Artist retains copyright.

Artist issues 10% participation in defined licensing income for five years.

Supporters finance the next production.

Revenue is distributed automatically.

After five years, the participation expires.

This is fundamentally different from selling the creator’s identity or entire copyright.

Any such financing instrument must undergo applicable financial, securities, consumer-protection and copyright-law review before issuance.

48. Patronage Reimagined

Not every creative transaction needs to be an investment.

Human Origin can also support patronage.

A supporter could finance:

an album;

photography expedition;

documentary;

painting series;

public installation;

book;

or cultural archive.

In return the supporter may receive:

early access;

limited editions;

event invitations;

acknowledgement;

collector status;

or other clearly defined non-financial benefits.

This creates direct human-to-human cultural funding.

49. Governments and the Human Creative Economy

Governments have a strategic interest in preserving human culture.

A Human Origin infrastructure could support:

national cultural archives;

arts grants;

museum provenance;

public broadcasting;

documentary photography;

historical records;

education;

cultural diplomacy;

and preservation of national cultural heritage.

A government could choose to require Human Origin provenance for certain cultural procurement without requiring the creator to expose unnecessary personal information.

50. Public Procurement of Human Creativity

Governments and institutions could deliberately purchase verified human-created works.

Examples:

human-created photography for public buildings;

human-performed music for cultural events;

human-created illustrations for schools;

human-written literature;

human-created public art;

human-directed documentary work.

AI remains available where appropriate.

But public institutions gain a transparent mechanism to preserve a market for human culture.

51. Human Origin and Journalism

The same technology has another major application:

VERIFIED HUMAN WITNESS

A journalist’s camera could establish:

physical sensor capture;

time;

device provenance;

human authorization;

editing history.

The objective is not to claim that every photograph is truthful.

Context can still be misleading.

But the system can establish:

this image originated from a physical capture rather than being generated from nothing by an AI model.

That distinction will become increasingly important.

52. Human Origin and Democratic Information

This connects directly with the election-integrity project.

Election infrastructure needs not only secure ballots.

Democracy also needs trusted information.

During elections, voters may encounter:

synthetic candidates;

fake speeches;

AI-generated photographs;

voice clones;

fabricated events;

and manipulated video.

A national Human Origin infrastructure could enable politicians, journalists, election authorities and citizens to verify provenance without giving governments the power to censor content.

The system should tell people:

what evidence exists about origin.

It should not decide:

what political opinion is true.

53. Connection to Election Integrity

The full architecture therefore becomes:

HUMAN PRESENCE

Who is acting?

↓

HUMAN AUTHORITY

What are they permitted to do?

↓

SECRET DEMOCRATIC AUTHORIZATION

Is this citizen eligible to vote?

↓

ANONYMOUS BALLOT

What did the citizen choose?

Identity is cryptographically separated.

↓

HUMAN ORIGIN INFORMATION

Which campaign media originated from humans, cameras or AI?

↓

PUBLIC AUDIT

Can institutions and citizens independently verify the evidence?

54. CIRAS Sovereign Hardware Family

The project can ultimately create a family of privacy-first devices.

CIRAS Ring

Presence and authorization.

CIRAS Watch

Biometric verification, approval interface and creator signatures.

CIRAS Bracelet

Advanced biosignal and liveness research.

CIRAS Secure Phone

Identity, sovereign AI, creator wallet and rights management.

CIRAS Creator Camera

Secure human-origin photography and video capture.

CIRAS Creator Audio Interface

Secure human-performance and recording provenance.

CIRAS Studio Node

Protected rights registration, storage, editing provenance and QFS settlement.

All should follow CIRAS’s existing research direction toward auditable, secure and reduced-emission hardware. CIRAS’s Post-Sovereign Cybersecurity Infrastructure project already identifies transparent computing architecture, post-quantum security and reduced electromagnetic emissions as research objectives.

55. Privacy: Proof Without Surveillance

Human Origin must not become continuous monitoring of artists.

An artist should never have to surrender:

every brush stroke;

every location;

every conversation;

every physiological signal;

or every draft

to prove authorship.

Evidence should be voluntary and proportionate.

The guiding principle remains:

Never collect what does not need to exist.

56. Human-Origin Data Architecture

The preferred architecture is:

raw creative work

→ creator-controlled storage

raw biometric

→ local device only

private identity

→ QSI / sovereign identity layer

public creator identity

→ optional professional credential

provenance

→ C2PA-compatible manifest

rights

→ Creator Rights Object

commercial licences

→ QFS transactions

payments

→ QFS settlement

public verification

→ Web B(eyond) proof interface

57. Post-Quantum Creative Archives

Creative intellectual property can remain economically valuable for a century or longer.

It therefore requires long-term cryptographic planning.

A photographic archive created in 2026 may remain culturally important in 2126.

CIRAS should research:

post-quantum signatures;

signature migration;

hash migration;

long-term timestamp preservation;

multi-generation key replacement;

museum-grade digital preservation;

and cryptographic archival continuity.

Human culture deserves the same long-term security thinking as financial infrastructure.

58. Web B(eyond) as a Cultural Trust Network

Web B(eyond) already proposes connecting identity, distributed infrastructure, digital objects, governance and scientific validation.

Human Origin adds:

CULTURE

to that architecture.

Web B(eyond) can therefore evolve from a digital nervous system for sovereign economic infrastructure into a network capable of supporting:

human identity;

human knowledge;

human creativity;

human cultural memory;

and human economic agency.

59. The Human Creative Economy

The long-term objective is a new economic model.

Today:

Platform owns distribution.

Platform owns user relationship.

Platform controls algorithm.

AI company controls model.

Creator competes with infinite synthetic supply.

The proposed model becomes:

Human owns identity.

Human creates work.

Human Origin establishes provenance.

Human controls rights.

AI requests permission.

QFS executes licences.

Web B(eyond) carries trust.

AI agents work for the creator.

Revenue returns directly to creators and collaborators.

60. What This Model Does Not Promise

Scientific credibility requires boundaries.

CIRAS should not claim that:

blockchain proves copyright;

a biometric proves authorship by itself;

AI-generated content can always be detected;

tokenization automatically transfers legal rights;

every AI-training use legally requires payment in every jurisdiction;

every creative work can be securely monitored;

or technology can eliminate copyright disputes.

Instead CIRAS can build something more defensible:

Better evidence.

Better attribution.

Better licensing.

Better human control.

Better settlement.

Better provenance.

Better economic choices for creators.

61. New CIRAS Research Workstreams

Workstream 1 — HEARTSPACE Human Presence Lab

Biometric identity, liveness and privacy.

Workstream 2 — Sovereign Identity & AI Authority

Human-controlled AI mandates.

Workstream 3 — Human Origin Protocol

Technical standards for proving creation provenance.

Workstream 4 — Secure Creator Hardware

Camera, audio, phone and studio hardware.

Workstream 5 — Content Credentials Integration

Interoperability with C2PA and other provenance systems.

Workstream 6 — Creative Rights Object

Machine-readable representation of IP rights and licences.

Workstream 7 — QFS Creative Asset Tokenization

Tokenized licences, royalty rights, editions and cultural assets.

Workstream 8 — AI Training Licence Market

A rights-cleared Human Creative Corpus.

Workstream 9 — Human Creativity Royalty Pool

Collective AI and media licensing.

Workstream 10 — Sovereign Creator AI

AI agents that economically represent artists under limited mandates.

Workstream 11 — Human-Origin Marketplace

Verified human-created cultural commerce.

Workstream 12 — Journalism & Authentic Media

Verified physical capture and provenance.

Workstream 13 — Election Information Integrity

Human and AI provenance in democratic communications.

Workstream 14 — Election Eligibility & Anonymous Voting

Biometric-protected eligibility separated from ballot identity.

Workstream 15 — Law, Copyright & Ethics

International legal mapping.

Workstream 16 — Independent Red-Team Programme

Artists, hackers, cryptographers, rights lawyers and AI researchers attempt to defeat the system.

62. Pilot I — Krista Kim / HEARTSPACE

HEARTSPACE becomes the philosophical origin of the Human Sovereignty project.

Participants experience:

my biological presence creates a unique digital response.

The exhibition could then introduce:

my presence can authorize technology.

And subsequently:

my human creation can carry sovereign rights.

This creates a cultural bridge between abstract cybersecurity and embodied human experience.

63. Pilot II — 100 Human Creators

CIRAS selects:

20 photographers;

20 musicians;

20 visual artists;

20 writers/designers;

20 filmmakers or multidisciplinary creators.

Each receives a Human Origin identity.

Their works enter an experimental provenance and rights environment.

Researchers study:

usability;

false claims;

collaboration;

privacy;

AI disclosure;

licensing;

payments;

and consumer response.

64. Pilot III — Human-Origin Photography

Professional photographers test secure camera capture.

A buyer sees:

HUMAN ORIGIN VERIFIED

and can examine:

camera provenance;

creator signature;

edit history;

AI-use statement;

licence;

ownership;

and available commercial rights.

This may become one of the clearest first commercial applications.

65. Pilot IV — Human-Origin Music

Musicians record a collaborative album.

Every contribution is registered.

Royalty splits are agreed before publication.

Commercial licences settle through QFS.

Selected tracks enter an opt-in AI-training corpus.

AI developers pay for access.

Creators receive transparent distributions.

The experiment tests whether a new machine-readable rights economy can materially reduce friction between AI development and musicians.

66. Pilot V — Human Creativity Royalty Pool

A larger group of creators voluntarily contributes works.

AI laboratories receive legally defined dataset licences.

QFS records payments.

Researchers compare:

creator income;

licensing cost;

administrative cost;

data quality;

model utility;

and creator satisfaction

against conventional approaches.

67. Pilot VI — Government Cultural Archive

A participating government or cultural institution uses Human Origin to preserve:

photography;

music;

oral history;

film;

art;

and cultural heritage.

The objective is permanent provenance and rights preservation.

Not surveillance.

68. Pilot VII — Election Information Integrity

During a non-binding democratic simulation:

official election communications use verified provenance;

journalists test human-origin camera capture;

synthetic campaign material is labelled;

candidate communications receive identity credentials;

and citizens can independently inspect provenance.

This operates separately from the anonymous voting system.

69. The Economic Objective

The purpose of the Human Origin layer is not merely technical.

It has a social objective:

HUMAN CREATIVITY MUST REMAIN ECONOMICALLY POSSIBLE.

Artificial intelligence should increase human capability.

It should not require society to accept a future in which professional photography, composition, illustration, writing, performance and filmmaking become economically impossible because synthetic production approaches zero marginal cost.

The answer is not necessarily to prevent AI from creating.

The answer may be to give human creation a different economic architecture.

70. Human Work Becomes Verifiable Scarcity

AI can create enormous quantities of synthetic content.

It cannot retroactively become:

the photographer who physically witnessed an event;

the singer who stood in the studio;

the painter who created the canvas;

the musician who performed the instrument;

the journalist who stood at the scene;

the human consciousness that experienced the moment.

That becomes the new scarcity.

Not scarcity of files.

Scarcity of authentic human experience.

71. CIRAS Human Origin Principle

The project can be summarized by a simple doctrine:

A human should own their identity.

A human should control their AI.

A human should be able to prove their creative origin.

A human creator should be able to determine how their work is licensed.

AI use should be disclosed.

AI authority should be bounded.

Creative rights should be machine-readable but legally grounded.

Tokenization should represent rights, not pretend to create them.

The creator should remain connected to the economic life of the work.

Privacy should be structural.

Democratic identity should never expose the secret ballot.

Technology should strengthen human agency rather than erase it.

72. From Proof of Human to Proof of Humanity

The larger CIRAS vision now becomes clearer.

The question of the AI age is not simply:

Can we build more powerful artificial intelligence?

It is:

How do we preserve the economic, political and cultural sovereignty of humans in a civilization increasingly operated by intelligent machines?

CIRAS Sovereign Presence establishes that a human is present.

Sovereign Authority establishes what that human has authorized.

Human Origin establishes what humanity has created.

Parisii QFS gives those rights economic and transactional infrastructure.

Web B(eyond) gives them an interoperable sovereign digital environment.

HEARTSPACE provides the artistic and philosophical starting point:

the living human remains at the center of the technological experience.

The combined vision therefore becomes:

PROOF OF HUMAN PRESENCE

Who is here?

↓

PROOF OF HUMAN AUTHORITY

What have they authorized?

↓

PROOF OF HUMAN CREATION

What did they create?

↓

PROOF OF HUMAN RIGHTS

Who controls the economic and moral rights?

↓

PROOF OF LICENCE

Who has permission to use them?

↓

PROOF OF AI AUTHORITY

What may the AI do?

↓

PROOF OF TRANSACTION

Who was paid?

↓

PROOF OF PROVENANCE

Can the history be independently examined?

73. Final Vision

Human beings should not have to compete with artificial intelligence on artificial intelligence’s terms.

AI can generate faster.

AI can copy patterns faster.

AI can produce more.

That does not make human creativity obsolete.

It makes human origin more important.

CIRAS can research an infrastructure in which the digital economy finally recognizes this distinction.

A photographer’s image can say:

A human witnessed this.

A song can say:

Humans composed and performed this.

A painting can say:

A human created this.

A film can say:

These scenes were physically captured and these scenes were synthetic.

A creator can say:

You may enjoy my work.

You may license my work.

You may train your AI on my work under these conditions.

You may not take ownership of my identity.

You may not authorize yourself.

And when economic value is created from that human work:

the human creator participates in the value.

This is the proposed foundation of the:

CIRAS SOVEREIGN HUMAN ECONOMY

A digital civilization in which human identity remains sovereign, artificial intelligence remains accountable, democratic choice remains private, human creativity remains distinguishable, and creators remain economically connected to the culture they produce.