[Paper Review] Human-Centric Artificial Intelligence Architecture for Industry 5.0 Applications

Jože M. Rožanec, Inna Novalija+18 more|arXiv (Cornell University)|Mar 21, 2022

Digital Transformation in Industry26 citations

TL;DR

The paper proposes a human-centric AI architecture for Industry 5.0 that integrates active learning, explainable AI, simulated reality, decision-making, and user feedback, mapped to the BDVA reference architecture, and validated on three real-world use cases.

ABSTRACT

Human-centricity is the core value behind the evolution of manufacturing towards Industry 5.0. Nevertheless, there is a lack of architecture that considers safety, trustworthiness, and human-centricity at its core. Therefore, we propose an architecture that integrates Artificial Intelligence (Active Learning, Forecasting, Explainable Artificial Intelligence), simulated reality, decision-making, and users' feedback, focusing on synergies between humans and machines. Furthermore, we align the proposed architecture with the Big Data Value Association Reference Architecture Model. Finally, we validate it on three use cases from real-world case studies.

Motivation & Objective

• Address the lack of an architecture that embeds safety, trustworthiness, and human-centricity at its core for Industry 5.0.
• Develop an architecture that enables effective human–machine synergies in manufacturing.
• Align the architecture with established reference architectures (BDVA) and ISA/ISSF frameworks for interoperability and standards.
• Validate the proposed architecture through three real-world industrial use cases.

Proposed method

• Propose an architecture that integrates active learning, forecasting, explainable AI, simulated reality, decision-making, and user feedback.
• Map architecture modules to the BDVA reference architecture model and ISSF framework to ensure compatibility.
• Discuss ethical governance, FAST principles, and security considerations as building blocks of the architecture.
• Validate the architecture across three real-world manufacturing use cases to illustrate applicability and interoperability.

Experimental results

Research questions

• RQ1 What is a viable architecture to realize human-centric AI in Industry 5.0 that ensures safety, trust, and human-in-the-loop collaboration?
• RQ2 How can the BDVA reference architecture be used to structure and align such an architecture with existing standards and frameworks?
• RQ3 Which enabling technologies (active learning, XAI, simulated reality, conversational interfaces, security) are essential for the proposed architecture?
• RQ4 How does the architecture perform and prove its value across three real-world Industry 5.0 use cases?

Key findings

• The architecture integrates AI components with human-in-the-loop to foster synergies between humans and machines.
• The design aligns with BDVA reference architecture and ISSF framework, illustrating interoperability and standards compliance.
• Validation on three real-world use cases demonstrates practical applicability in Industry 5.0 contexts.
• The work emphasizes safety, trust, ethics, and governance (FAST principles) as core governance aspects of human-centric AI in manufacturing.
• The study outlines enabling technologies (AL, XAI, simulated reality, conversational interfaces, security) as foundational elements of the architecture.