Count Learning Value, Not AI Tools: The Measure of Responsible Digital Transformation

Higher education has entered a dangerous stage of artificial-intelligence adoption: the tools are increasingly visible, but their educational value is still uncertain. The Digital Education Council's 2026 global survey brings unusual scale to this concern. It draws on 45,398 responses from students and faculty across thirty-five countries. Only fifteen per cent of students said AI was integrated into many of their courses, while forty-three per cent encountered it in a few and another forty-three per cent had experienced no course integration. Yet even where AI was present, only five per cent said it had transformed how they learned. Twenty-eight per cent reported stronger understanding, forty-two per cent found it only somewhat helpful and twenty-four per cent saw no clear learning value.

Those figures should change the question asked in every boardroom, college, church and training organisation. The first question must not be, ‘How many AI tools have we adopted?’ It must be, ‘What are our learners now able to understand, judge, create and do that they could not do before?’ Technology adoption is an activity. Educational transformation is a demonstrated human outcome.

Adoption without learning value is not transformation

Institutions can mistake movement for progress. A lecturer may generate more lesson material, a student may submit more polished work and an administrator may produce documents faster, while understanding, originality and professional judgement remain unchanged. AI makes output easier to count, but the most important outcomes—discernment, transfer, ethical judgement, practical competence and mature responsibility—remain human and require better evidence.

For African theological colleges, this means that an AI initiative should begin with curriculum outcomes. If the outcome is biblical interpretation, learners must interpret faithfully. If it is pastoral care, they must listen, judge and respond responsibly. If it is governance, they must weigh evidence, explain a decision and accept accountability. AI may support these processes, but a fluent output cannot be accepted as proof that formation has occurred.

Training attendance is not the same as teaching readiness

The same global survey reveals a credibility gap. Sixty-four per cent of faculty said they had participated in AI-literacy training, yet only twenty-nine per cent of students believed their instructors were well equipped to guide AI use. Microsoft’s 2026 education report points in the same direction: sixty-six per cent of educators and fifty-two per cent of students want AI training monthly or quarterly. A single workshop may create awareness, but it rarely changes curriculum, assessment or classroom judgement.

Professional development should therefore become a continuing practice rather than an occasional event. Lecturers need opportunities to redesign one real assignment, test a disclosure statement, examine an AI-generated answer, compare student work before and after intervention and discuss the results with colleagues. The unit of progress is not the certificate issued to a lecturer. It is the improved learning experience that students can recognise.

Assessment must recover competence and trust

Only twenty-eight per cent of students in the Digital Education Council survey believed that most or many of their assessments reflected the work, skills and judgement they expected to need in an AI-enabled workplace. Sixty per cent also worried that peers might use AI for unfair advantage. These are not separate concerns. When assessment does not resemble meaningful practice and its rules are unclear, students lose confidence both in relevance and fairness.

Competency-based theological education offers a constructive response. Assessment can combine an AI-use declaration, annotated sources, an oral defence, observed ministry practice, contextual application and reflection on what the learner changed after receiving feedback. Some tasks should be AI-free to reveal foundational competence; others should require responsible AI use because verification and tool judgement are themselves emerging professional capabilities. The governing principle is simple: every assessment should make the intended competence visible.

Theological education must test fluency against faithfulness

Recent theological-education scholarship has begun to move beyond general excitement or fear. Faith Seeking Prompting argues for pedagogical and institutional models rather than scattered individual experimentation. Research on doctrinal flattening adds a sharper warning: language models can produce confident religious explanations while blurring distinctions among Christian traditions. For a Pentecostal institution, polished theological language is therefore not sufficient evidence of biblical or doctrinal faithfulness.

Students should learn to test AI-generated theology against Scripture, responsible exegesis, historical Christian teaching, Pentecostal doctrine, trusted scholarship and lived ministry context. They should be able to identify what a generated answer omits about the work of the Holy Spirit, spiritual gifts, sanctification, healing, mission, communal discernment or African Christian experience. This is not resistance to knowledge. It is the Pentecostal practice of testing, discerning and holding fast to what is good.

The Global South needs its own learning-value test

Africa must not assess educational progress through imported technology counts. A Pan-African ethical-AI curriculum proposal rightly connects technological capability with critical thinking, cultural context, Ubuntu-informed relational ethics and social justice. The forthcoming KNUST E-Learning International Conference similarly places inclusion, equity and ethics inside the discussion of AI and emerging technologies. These emphases matter because a tool that works in a highly resourced university may fail a learner studying on an ordinary phone, purchasing data in small amounts or learning across several languages.

A credible African learning-value test should ask whether the design is accessible on modest devices, grounded in approved curriculum, relevant to local realities, transparent about AI involvement, protected by human oversight and capable of producing observable competence. It should also ask who is excluded, whose language and knowledge are missing and whether the institution can sustain the system after external enthusiasm or funding declines.

Build a learning culture, not a tool collection

For theological colleges, church education, church leadership development and public-service training, the practical next step is a ninety-day learning-value pilot. Select a small number of courses, establish baseline evidence, redesign specific activities and measure student understanding, completion, feedback quality, lecturer workload, integrity, accessibility and contextual application. Listen to students and trainers before expanding. Let governance review both benefits and risks.

The same principle should govern digital ministry and ethical enterprise. A course, ebook, membership, template or AI-enabled learning asset is valuable only when it helps people become more capable. We should not sell volume that has not been transformed into wisdom. We should create resources that carry genuine expertise, invite practice and produce measurable improvement. The future does not belong to institutions with the largest collection of AI tools. It belongs to communities that can show what their people are learning, how their judgement is maturing and why their transformation can be trusted.