Implementation
The Liccium Platform
The FAIA attribution framework is integrated into the Liccium platform as a plugin. This enables creators, researchers, publishers, and platforms to declare AI involvement directly within their content authentication workflows. Designed to be intuitive yet technically robust, the plugin supports transparency, compliance, and interoperability by making AI contributions to content creation verifiable and persistently linked to the original work.
How does it work?
1. Select the Type of AI Involvement
Users can choose from a set of standardised Flags that reflect the degree and nature of AI’s contribution. These include the following categories:
HCC
Human-Created Content
Content created and edited exclusively by humans. While digital tools may be used, no AI systems are involved at any stage of the creative or editorial process.
AAC
AI-Assisted Content
Content where a human remains the primary author but AI systems contributed during the process. This may include suggestions, generation of fragments, or refinement steps performed under human supervision and editorial control.
AIG
AI-Generated Content
Content generated predominantly or entirely by an AI system. The AI is the main creative agent. Human input may be limited to initiating a prompt or selecting from outputs, with minimal or no human editing or authorship applied.
These flags provide high-level signals of AI involvement. It is intended for use in content metadata, declarations, digital packaging, or registry records.
Be aware that the framework is still in development!
2. Provide Contextual Metadata
To increase transparency and support downstream processing, FAIA supports additional metadata describing what was done, who did it, and how:
a. Activity Type
Specifies the operation performed on or to the content. FAIA supports activity codes from:
IPTC (for image-related workflows)
STM (for publishing workflows)
FAIA (generic cross-media activity types such as
faia:Contribution,faia:Enhancement, orfaia:Generation)
b. Actor
Each declared activity must be attributed to one of:
Human Actor– the operation was performed by a personAI Actor– the operation was performed by an automated system
c. System Metadata (only if the actor is AI)
If the activity was performed by an AI system, the following fields may be included to support reproducibility and audit:
Tool – Name of the system or interface (e.g.
"ChatGPT")Model – Underlying model (e.g.
"GPT-4o")Version – Model or tool version (e.g.
"4.0")Provider – The organization offering the AI system (e.g.
"OpenAI")
Be aware that the framework is still in development!
3. Bind to the ISCC Code
All FAIA declarations are cryptographically bound to the International Standard Content Code (ISCC), a content-derived identifier and ISO standard (ISO 24138). This ensures that attribution metadata remains uniquely and persistently associated with the content – even if the asset is modified and redistributed.
4. Generate Verifiable, Timestamped Declarations
Every declaration is digitally signed using Verifiable Credential (VC) or DID keys of the creator. The declaration is stored as a tamper-evident JSON-LD file (nanopublication), including a trusted timestamp. This guarantees the authenticity and integrity of the AI attribution data.
5. Register in the Liccium Registry
The declarations are published to a public, decentralised registry operated by Liccium, making them:
Publicly resolvable using content identifiers like the ISCC
Searchable for researchers, platforms, and compliance tools
Auditable under regulatory regimes such as the EU AI Act (e.g. Article 50 disclosure obligations)
This registry ensures lasting transparency and verifiability in content attribution and provenance.
Last updated