Abstract
Innovation scholars need to apply theory and practice to manage changes. Generative artificial intelligence (AI) has introduced a legitimacy crisis into doctoral education. Text can now be produced faster than it can be evaluated, empirical patterns can be simulated without data, and methodological reasoning can be obscured behind opaque automation. Regulators increasingly question whether dissertations still demonstrate the independent causal reasoning traditionally expected of doctoral graduates. Contributing to innovation practice and its methodologies, this paper proposes a forward-looking solution grounded in the Causal Theoretical Twin Architecture (CTTA), a general framework for multi-scale causal modelling. We argue that CTTA embeds the student’s intellectual “fingerprints” into the structure of the research itself through the design of theoretical, data, empirical, and virtual twins. These twins require human judgement, causal justification, and structural design that cannot be credibly delegated to AI tools. The paper develops a practice based, real world oriented CTTA-based workflow for doctoral research, a rubric for evaluating structural originality, and a positive vision for strengthening doctoral quality control in the post-AI era.