Why Human Layer? AI Assistance, Human Authorship and Accountability

I use AI in my creative work. That experience is part of why I want to build spaces where human judgment and responsibility remain visible.

AI helps me turn ideas into drafts, images, structures, questions and connections between projects. Human Layer's own website was developed with AI assistance. Its generated illustrations are labelled, and the person responsible for selecting and publishing them is identified.

The practical question is what a person contributes when powerful tools help produce a work—and what they take responsibility for.

A label saying “AI-generated” answers only part of that question. So does “human-made”. Neither tells us whether someone checked the sources, understood the argument, made meaningful editorial decisions or will respond when a mistake is found.

AI assistance and meaningful human authorship

AI can help someone translate an idea, organise a difficult project or communicate through an accessibility barrier. It also introduces decisions that cannot simply be delegated away.

What should be kept? What needs verification? Does an image communicate the intended idea or create a misleading impression? Does a fluent explanation actually make sense?

Choosing to publish brings responsibility for those decisions. Using AI does not automatically remove human authorship. Attaching a name to an output does not automatically establish it either.

Different kinds of work make different promises. A conceptual illustration and a historical documentary require different disclosures and evidence. Where assistance becomes substitution remains a question we need to work through openly.

What are we asking people to trust?

When we watch a video, several questions can become tangled together. Did this event happen? Was the footage recorded or generated? Who wrote the explanation? Where did the claims come from? Who can answer for the result?

Media provenance concerns the origin and history of the material. Information provenance concerns the sources and evidence behind its claims.

A real camera can record a false statement. Authentic archival footage can be placed in a misleading context. A bibliography may fail to support the central argument. Making origins visible helps us ask better questions; it does not remove the need to ask them.

For factual work, viewers should be able to inspect references, connect claims to evidence and find corrections. For creative work, they should understand how it was made without confusing illustration with documented reality.

Human attention belongs in the design

These concerns connect with my longer exploration of perception and modern information culture. What reaches us becomes part of the environment in which we form judgments.

A first publication should have a route to interested people without requiring months of follower collection. Discovery could follow interests, understandable selection rules and voluntary recommendations. Following a person can be valuable; it should not be a prerequisite for being heard.

The financial structure must support the purpose

Moderation, development and support deserve fair compensation. At the same time, Human Layer's founding requirements include binding limits on private enrichment. After operating costs, necessary development and reasonable reserves, remaining surplus should benefit charitable organisations.

Those limits must be secured through ownership, financing and decision-making rules. No such legal structure has yet been established. Making it enforceable is part of the work ahead. Read the Fair Attention & Shared Benefit principles.

An idea in public

Human Layer currently has a public website, a working Constitution and proposals for a future platform. It does not yet have a video service, verification system or operating moderation team.

We need people who can challenge and improve the idea: creators, developers, journalists, historians and people experienced in provenance, moderation and governance.

AI can help us create more. Human Layer asks how we can remain accountable for what we create and respect the people receiving it.

Read Constitution V0.1 · Explore the platform vision

Discuss the idea or get involved · More from the Human Layer Journal

By Paul-Lukas Hertz. Developed with ChatGPT assistance from the Human Layer founding principles. Human Layer is currently in Phase Zero; the platform features discussed here are proposals, not operating services.