Evaluating generative AI tools for visual communication design using the CoCoSo method under interval valued...

Area of Science:

• Artificial Intelligence
• Visual Communication
• Decision Science

Background:

• Generative AI is transforming visual communication by automating creative processes.
• Evaluating generative AI tools requires robust decision-making frameworks.
• Existing methods may not fully capture expert uncertainty in assessments.

Purpose of the Study:

• To propose a multi-criteria decision-making (MCDM) approach for analyzing generative AI tool efficiency.
• To introduce an interval-valued spherical fuzzy (IVSF) framework combined with the CoCoSo method.
• To provide a practical evaluation of prevalent generative AI platforms.

Main Methods:

• Developed a novel IVSF-CoCoSo approach to handle expert uncertainty in evaluations.
• Applied the framework to a case study assessing ten generative AI platforms across seven criteria.
• Included creativity, cost, customization, and integration capability in the assessment.

Main Results:

• The IVSF-CoCoSo approach identified DeepArt as the most effective generative AI tool.
• DeepArt excelled in creativity improvement, output quality, and customization.
• Sensitivity analysis confirmed the proposed model's reliability and flexibility.

Conclusions:

• The IVSF-CoCoSo method offers a reliable and flexible approach for evaluating generative AI tools.
• DeepArt is recommended for designers and researchers seeking advanced visual communication solutions.
• This study provides practical insights for implementing generative AI in creative industries.