Blockchain and NFTs for Trusted Ownership, Trading, and Access of AI Models

Ammar Battah, Mohammad Madine, Ibrar Yaqoob, Khaled Salah, Haya R. Hasan, Raja Jayaraman

Research output: Contribution to journalArticlepeer-review

16 Scopus citations

Abstract

The demand for high-quality Artificial Intelligence (AI) models is ever-increasing in this digital era. However, most of the existing methods leveraged for managing the ownership, trading, and access of AI models fall short of providing traceability, transparency, audit, security, and trustful features. In this paper, we propose a solution based on blockchain and Non-fungible Tokens (NFTs) to manage ownership rights and exchange of AI models in a transparent, traceable, auditable, secure, and trustworthy manner. Smart contracts are employed to enforce ownership, ease of access, and exchange policies for the unique NFT linked to an AI model. We use decentralized storage of the InterPlanetary File System (IPFS) and proxy re-encryption oracles to securely fetch, store, and share data related to AI models. We present algorithms along with their implementation, testing, and validation details. The proposed solution is evaluated using cost and security analyses to show its affordability and resiliency against security threats and attacks. All smart contract codes are made publicly available on GitHub.

Original languageBritish English
Pages (from-to)112230-112249
Number of pages20
JournalIEEE Access
Volume10
DOIs
StatePublished - 2022

Keywords

  • Blockchain
  • decentralized storage
  • non-fungible tokens (NFTs)
  • oracles
  • provenance
  • proxy re-encryption
  • smart contracts

Fingerprint

Dive into the research topics of 'Blockchain and NFTs for Trusted Ownership, Trading, and Access of AI Models'. Together they form a unique fingerprint.

Cite this