Designing an AI-driven framework for enhancing interaction practices in industrial systems
摘要
This study presents a novel framework called the Co-Creative Framework for Interaction (COFI), which integrates human and machine intelligence to facilitate structured collaboration between people and machines, enhancing decision-making and workflow design. The framework was assessed through a mixed-method evaluation that included quantitative analysis of 4820 process records, a controlled user study with 1000 professionals, and cost–time assessments using a financial dataset. The performance was compared across three models: human-only, machine-only, and co-creative (combination of human and machines). The novelty of the framework introduces a composite evaluation structure (creativity, efficiency, adaptability, and user satisfaction) tailored to decision-making in safety–critical domains, which differentiates it from purely usability-oriented models. The results showed that the proposed COFI model consistently achieved higher creativity scores (80%), stronger user satisfaction (85%), adaptability (70%) and model accuracy (90%) while maintaining efficiency and reducing operational costs. The statistical analysis confirmed the significance of these outcomes, demonstrating COFI’s ability to balance innovation with practicality. Unlike human-only approaches, which struggle with scalability, and machine-only models, which sacrifice engagement, COFI integrates the strengths of both.