Human Layer — Constitution V0.1

Human Layer · About · Constitution · Impressum · Privacy

Working constitution · 30 September 2026

This is the founding text of Human Layer, published for discussion. It sets our direction; it does not claim that a video platform or verification service already exists.

Working Premise

The internet is entering an era in which synthetic content can be produced at effectively unlimited scale.

Human attention cannot.

Producing another article, image, voice or video can become extraordinarily cheap, while evaluating, verifying, storing, transmitting and consuming that information continues to consume real resources.

The Problem

Generative AI can enormously increase human creativity and agency.

At the same time, it allows content production to be industrialized without a corresponding increase in human knowledge, creativity or responsibility.

Low-value synthetic material can be converted into real economic value through advertising, subscriptions, affiliate revenue and algorithmic attention.

The production cost may approach zero.

The real costs do not.

Computation, energy, storage, bandwidth and infrastructure are consumed.

Most importantly, human attention remains scarce.

At its least harmful, this produces noise.

At its worst, it produces misinformation, fabricated history, synthetic testimony, false expertise and an information environment in which documentation becomes increasingly difficult to distinguish from simulation.

The problem is therefore not simply that AI can generate bad content.

The problem is an information economy capable of rewarding enormous quantities of content regardless of whether that content contributes meaningful informational, cultural or creative value.

What We Want to Protect

We are NOT attempting to create an internet without AI.

We want to preserve:

Human Agency.
Human Authorship.
Human Accountability.
Human Creativity.
Provenance.
Traceable Sources.

People should be able to know when another human being meaningfully stands behind what they are seeing.

Human Agency

This is the central principle.

A human author makes choices.

A human researcher evaluates evidence.

A human filmmaker decides what deserves to be shown.

A human creator can be questioned, challenged, corrected and held responsible.

AI can dramatically extend these capabilities.

Human Layer therefore does NOT fundamentally distinguish between:

AI vs. Human

but rather between:

AI serving human agency

and

automated systems replacing human authorship while competing for human attention.

AI assistance is not automatically the problem.

The disappearance of identifiable human authorship and responsibility is.

Provenance

Human Layer should eventually explore two different forms of provenance.

Media Provenance

Where did this material come from?

Possible future mechanisms include:

  • C2PA / Content Credentials
  • creator verification
  • original production material
  • production histories
  • random audits
  • transparent declarations

Information Provenance

Where did this claim come from?

Human origin does not guarantee truth.

A genuine human can produce propaganda, misinformation or fabricated history just as easily as a machine.

Therefore factual content should make its sources visible and accessible.

Potential future concepts include Source Cards for documentaries, history, journalism, science and other factual material.

Provenance Before Detection

The project should NOT be based primarily on attempting to perfectly detect AI-generated content.

Generation technology evolves.

Detection technology fails.

False positives can harm legitimate creators.

The more useful question is:

Not only: “Can we prove this was generated by AI?”

but:

“Can we establish where this came from?”

The objective is not technological certainty.

It is:

Accountability by design.

What Human Layer Is NOT

Human Layer is:

not anti-AI.

It is not technological nostalgia.

It does not assume human-created content is automatically valuable.

It does not assume AI-assisted work is automatically inauthentic.

It should not become an arbitrary technological purity test.

Examples that may be compatible with human authorship include:

  • transcription
  • translation assistance
  • noise reduction
  • restoration
  • accessibility tools
  • research assistance
  • editing assistance

The difficult boundary is where:

assistance becomes substitution.

Defining this boundary openly is part of the project.

Long-Term Concept

If sufficient community and expertise emerge, Human Layer may eventually become a curated video platform centered around:

  • identifiable human creators;
  • meaningful human authorship;
  • transparent AI assistance;
  • media provenance;
  • information provenance;
  • sourcing for factual claims;
  • human moderation;
  • creator accountability;
  • protection against automated content farming.

PeerTube is currently an interesting technical foundation.

C2PA / Content Credentials are interesting for provenance.

Neither technology is currently assumed to be the final solution.

Phase Zero Success Criterion

We are NOT measuring success by traffic.

The first important question is:

Who shows up?

Before building a technical video-platform MVP, we want to find approximately:

  • 5+ serious collaborators or contributors;
  • technically experienced people;
  • creators interested in publishing;
  • expertise outside our own bubble, ideally including areas such as journalism, history, provenance, moderation or Trust & Safety.

Until meaningful collaborators emerge:

DO NOT TURN THIS INTO A GIANT SOFTWARE PROJECT.

Website → discussion → people.

Only then:

prototype → platform.

Transparency principle — public implementation

We disclose which generative AI tools were used, how they contributed to the work and who takes human responsibility for the result.

Descriptions should distinguish concept development, generation, selection, editing and publication. Generated illustrations must not be presented as documentary evidence.

This principle was defined during the public website preparation and supplements the original Constitution V0.1 above.

Authorship and development

Human Layer was initiated by Paul-Lukas Hertz. Its founding text and website were developed in dialogue with ChatGPT, with AI assistance in drafting, structuring and implementation. Paul-Lukas Hertz sets the direction, reviews the work and takes responsibility for publication.

Discuss the constitution or offer a contribution

Human Layer · About · Constitution · Impressum · Privacy