CBTE Must Prove the Evidence Belongs to the Learner

A learner can now produce an elegant sermon manuscript, exegetical essay, ministry plan, reflective journal or case analysis with extensive assistance from generative artificial intelligence. The document may be useful. It may even be true. Yet a competence decision asks a different question: what does this evidence allow an assessor to conclude about the learner?

That distinction reaches the heart of Competency-Based Theological Education. CBTE is the international abbreviation for this approach and refer here principally to contextual, mentored formation in which learners demonstrate integrated knowing, being and doing for ministry. This is distinct from Kenya's general occupational Competency-Based Education and Training, commonly abbreviated CBET.

CBTE has often valued portfolios because ministry competence cannot be reduced to one written examination. But a portfolio is not self-authenticating. In an AI-rich environment, a programme must be able to defend the connection between the artefact, the learner and the competence claimed. The central issue is not whether AI touched the work; it is whether the resulting evidence still supports the credential's claim.

A peer-reviewed conceptual paper published online on 20 September 2026 argues that generative AI creates several different threats to assessment validity. It can change the competence being sampled, introduce irrelevant influences, weaken attribution of performance to the learner, limit transfer to other conditions or create harmful uses of a decision. The author proposes beginning with the…