AI Trustworthiness in Manufacturing: Challenges, Toolkits, and the Path to Industry 5.0

Area of Science:

• Manufacturing Technology
• Artificial Intelligence
• Industrial Systems Engineering

Background:

• Industry 4.0 is evolving towards Industry 5.0, emphasizing human-centric, sustainable, and resilient manufacturing.
• Artificial Intelligence (AI) is pivotal in this transition, enabling advancements in predictive maintenance, quality control, and supply chain optimization.
• Widespread AI adoption faces significant hurdles, including the "black box" problem, data biases, ethical considerations, and a lack of robust trustworthiness frameworks.

Purpose of the Study:

• To provide a comprehensive survey of Artificial Intelligence (AI) trustworthiness within the manufacturing sector.
• To examine the evolution of industrial paradigms and identify key barriers to AI adoption.
• To explore principles of AI trustworthiness, including transparency, fairness, robustness, and accountability.

Main Methods:

• Literature review of AI trustworthiness in manufacturing.
• Analysis of existing toolkits and methodologies for AI explainability, bias mitigation, and robustness.
• Examination of challenges across the AI pipeline, from data collection to model deployment.

Main Results:

• Key barriers to AI adoption in manufacturing identified, including model opacity and data integrity issues.
• Existing methodologies for enhancing AI transparency, fairness, and robustness are summarized.
• Challenges in the AI lifecycle, from data acquisition to deployment, are highlighted.

Conclusions:

• Fostering trust in AI systems is crucial for realizing the full potential of Industry 5.0.
• Actionable insights and research questions are provided to guide the development of ethical and reliable AI in manufacturing.
• The study aims to support researchers, practitioners, and policymakers in aligning AI advancement with societal values.