TAIRA-BSC - Trusting AI in Recruitment Applications through Blockchain Smart Contracts

Abstract

Artificial intelligence (AI) and blockchain technology (BCT) are considered two of the most trending and disruptive technologies. BCT, although commonly associated with cryptocurrencies, has shown a tremendous impact among many other distributed applications domains. BCT characteristics, such as the distribution of data storage among independent nodes and the use of consensus algorithms offer immutability and transparency and remove the need for a central authority making BCT trustworthy. However, decision-makers and stakeholders currently lack the confidence to overcome uncertainty related to AI technology, which affects the acceptance of AI technology in wider application domains, such as the recruitment process. Furthermore, current research literature does not adequately investigate the role of trust as an integral part of an AI-based recruitment application. Therefore, this paper aims to investigate how emerging BCT and AI technologies can improve decision making and stakeholder trust in a job recruitment system that is traditionally focused on just human expert decision-making. In this paper we propose the design of a new solution for trusting AI in recruitment applications through the use of Blockchain Smart Contracts (TAIRA-BSC). TAIRA-BSC integrates Blockchain Smart Contracts (BSC) with the Data Lake (DL), Machine Learning (ML) and AI technologies in our AI Recruitment Model (AIRM) architecture. TAIRA-BSC improves transparency and interoperability in the recruitment process while protecting sensitive job candidate data and ensures data integrity delivery and traceability in the recruiting process through a verifiable decentralized ledger, i.e., the blockchain and associated smart contracts. The paper presents a discussion on the state-of-the-art of integrating AI with BCT focusing on how BCT can be used to bridge trust concerns with AI systems. We also present a conceptual architecture TAIRA-BSC proof of concept that is developed to serve as a foundation for future studies focused on enhancing trust in AI applications through the integration of BCT.