From Automation to Coevolution: A Systematic Review of Human–AI Interaction Hierarchies and Professional Implications in the Construction Industry

Abstract

Artificial intelligence (AI) is rapidly reshaping the construction industry, creating increasing reliance on human–AI interaction across project processes. Although previous research has made significant progress in applying AI across construction workflows, existing literature reviews lack a systematic analysis of human–AI interaction patterns, coupling depths, and the implications for professional competence. To close this gap, this study aims to (1) systematize the hierarchical levels of human–AI interaction in construction applications; and (2) identify the professional competence requirements corresponding to different interaction levels. A Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA)-guided systematic review was conducted, covering 57 peer-reviewed studies from 2016 to 2025, complemented by bibliometric mapping and thematic synthesis to identify patterns of interaction and competence demand in construction contexts. The review proposes a four-layer hierarchical framework of human–AI interaction: AI automation (Layer 1), AI assistance (Layer 2), AI–human coordination (Layer 3), and adaptive coevolution (Layer 4). Layers 1–3 emerged from empirical analysis of construction practice, and Layer 4 represents an emergent stage informed by nascent construction technologies and cross-domain coevolution frameworks. Each layer reflects distinct coupling intensity, interaction dynamics, and competence requirements. Building on these insights, this study proposes a human–AI interaction literacy matrix that systematically maps distinct levels of interaction depth to evolving professional competence requirements across technical-operational, analytical-evaluative, collaborative-interactional, and reflective-adaptive dimensions. These findings provide actionable guidance for construction organizations to design role-differentiated training programs aligned with interaction depth, assess workforce readiness for AI integration at specific project phases, and develop career pathways as human–AI interaction evolves from automation-dominant to collaboration-intensive configurations. The framework redefines AI adoption as a competence-driven progression rather than a purely technological implementation.