Determine optimal designs and human dimensions of adaptive AI-centric lifecycles

Determine the optimal designs and human dimensions of continuous, adaptive software-development lifecycles for data-intensive and AI-enabled systems, including their collaboration, explainability, skills, ethics, legacy-integration, and security requirements.

Background

The paper presents a five-layer adaptive lifecycle framework based on coupled artifact versioning, explicit contracts, statistical gates, closed-loop control, and governance by construction. It identifies the framework as conceptual rather than empirically validated and emphasizes that the field still lacks evidence about which lifecycle designs are optimal and how organizational, ethical, human, and operational constraints should shape those designs. These unresolved dimensions are also reflected in the research agenda’s calls for design-science evaluation, sustainability analysis, human-factors research, and educational research.