ZAP AI Protocol

ZAP AI Protocol is an open source Web3 framework designed to facilitate the creation, management and monetization of decentralized Knowledge Assets for AI applications, to kickstart KnowledgeFi. ZAP AI Protocol empowers AI value-creators to connect up their expertise - whether this lies in data production, model training, app design or elsewhere - and enjoy transparent accounting and revenue shares. Under this system, each component is wrapped in an NFT or an ERC-3525 Semi-Fungible Token (SFT), allowing for easy, low-gas transfers of economic value between the components in real time, users interact with them.

The $ZAP token serves as an accounting unit as well as the core transaction currency of the protocol.

ZAP AI Protocol comprises 3 layers:

1. Application Layer

AI apps can be deployed on ZAP AI Protocol via the minting of an NFT or SFT, seamlessly linking to AI models and Knowledge Assets and allowing for interactions to be recorded as entries on the $KIP ledger.

2. Settlement Layer (Accounting Layer)

Connected apps, models and Knowledge Assets transfer value denominated in $ZAP between each other efficiently. Technical details here.

3. Ownership Layer

Data & Knowledge can be stored (and encrypted) in an on-chain location in such a way that they can be made accessible to selected AI models for retrieval-augmented generation (RAG) or training, but inaccessible to humans who may be tempted to pirate them.

Anyone may use ZAP AI Protocol to create, protect and secure their own Knowledge Assets on chain.

Below is an overview of ZAP AI Protocol functions:

Knowledge Asset creation

Minting of NFTs and SFTs.

Knowledge Asset trading (including fractionalised trading)

Order placement and payment and asset transfer. Fractionalising of Knowledge Asset SFTs for greater liquidity and portfolio construction by Knowledge Investors.

Network verification

Nodes check API uptime and report back to the ZAP AI DAO regarding broken links or malicious providers.

Automated revenue share

Automated split and sharing of revenues among app, model and dataset owners.

Knowledge Asset crowdfunding

App, model and dataset owners can demonstrate past revenue and seek funds for further expansion by selling a portion of their future revenues to investors.

Token Gating

Creator use of ZAP AI Protocol NFTs and SFTs to token-gate access to their own works.

Decentralised RAG

Using KIPZAP AI Protocol to carry out decentralised RAG, incorporating contributions from app developers, model trainers and data owners, ensuring fair payments to each.

Decentralised RLHF (experimental)

Using ZAP AI Protocol to organise & incentivise human reinforcement learning in AI model training.

Decentralised recording nodes (experimental)

Use of websocket access nodes and eventually possibly an independent app chain with its own RPC nodes and validators for the purpose of storing data.