Data or Design First? Rethinking the roles and challenges of designers in the era of data-centric Artificial Intelligence

Abstract

The shift from model-centric to data-centric artificial intelligence (AI) represents a paradigm change that demands active engagement from designers. Using a high-level literature review and the Data-Information-Knowledge-Wisdom (DIKW) framework, this study identifies five key challenges designers face in AI development: aligning AI with user needs, leveraging small yet high-quality user data, uncovering nontrivial and meaningful patterns, refining AI models through iterative usability testing, and envisioning robust data pipelines. These challenges underscore the critical role of human input in mitigating blind spots in AI systems and fostering practical, human-centered solutions. The results emphasize the transformative potential of collaborative intelligence—an active learning process between human designers and AI systems. This approach bridges the gap between abstract computational processes and real-world applications, empowering designers to drive innovation while ensuring ethical accountability.