What the CCL contributes — and how it compares.
Authorship in the age of AI is no longer binary.
The gap the CCL fills
A rich ecosystem of AI-transparency tools already exists.
Its most original piece is the Human-first label.
Comparison with other AI labels
| Tool | What it documents | Reader | Verifiable | Human focus |
|---|---|---|---|---|
| CCL | The creative process — AI vs human contribution, phase by phase | Human | No — by design | High |
| C2PA | Provenance & edit history of a file | Machine | Yes (crypto) | None |
| Model Cards / Datasheets | How a model or dataset was built | Human + machine | Partial | None |
| AID / DAISY | Which AI tools were used in a manuscript | Human | No | Low |
| CRediT | Contribution roles between co-authors | Human | No | High (people) |
| CC Signals | Permissions for AI training use | Human + machine | Flexible | Partial |
The CCL is the only one that combines a focus on the creative process, an explicit account of the irreplaceably human, and a reflective, pedagogical purpose.
Complement, not competitor
vs the EU AI Act
The Act obliges marking the machine's output; the CCL declares the author's process.
vs education guidance
UNESCO and the EU ask for reflection but give no artifact. The CCL is that artifact.
Why non-verifiable is a stance, not a flaw
The CCL is self-declared and cannot be cryptographically proven — and that is deliberate.
Aligned with the fab lab commons
Open, distributed, peer-based, documentation-as-practice.
Sources
EU AI Act — Article 50 · C2PA · Datasheets for Datasets · AID Framework · CRediT (NISO) · CC Signals · UNESCO AI Competency Framework