Search papers, labs, and topics across Lattice.
This paper addresses the lack of a consensus definition of human-centeredness in AI by developing a hierarchical framework comprising 26 attributes across four tiers: ethical foundations, usability, emotional/cognitive dimensions, and personalization. The framework was validated through practitioner input, emphasizing ethical considerations, usability, and emotional intelligence as key components. The resulting framework offers actionable guidance for AI practitioners to design AI systems that prioritize human well-being and societal values.
AI practitioners now have a validated, hierarchical framework of 26 attributes to guide the development of truly human-centered AI systems.
There is no consensus on what constitutes human-centeredness in AI, and existing frameworks lack empirical validation. This study addresses this gap by developing a hierarchical framework of 26 attributes of human-centeredness, validated through practitioner input. The framework prioritizes ethical foundations (e.g., fairness, transparency), usability, and emotional intelligence, organized into four tiers: ethical foundations, usability, emotional and cognitive dimensions, and personalization. By integrating theoretical insights with empirical data, this work offers actionable guidance for AI practitioners, promoting inclusive design, rigorous ethical standards, and iterative user feedback. The framework provides a robust foundation for creating AI systems that enhance human well-being and align with societal values. Future research should explore how these attributes evolve across cultural and industrial contexts, ensuring the framework remains relevant as AI technologies advance.