AI in Education Needs More Than Tool Experts

A lecturer knows how people learn but cannot explain what data an artificial-intelligence platform collects. A software specialist can configure the platform but has never designed a valid assessment. A senior leader approves a licence without knowing whether the system supports the curriculum. An ethics committee writes principles that never reach the classroom.

Every person may be competent within a particular role. The institution still lacks someone capable of joining the roles around an educational purpose.

Artificial intelligence in education is moving beyond scattered experimentation. It is entering degree programmes, teacher preparation, procurement, assessment, student support and institutional planning. That expansion creates a new professional responsibility: to translate between learning and technology without allowing either language to dominate the other.

The future does not belong merely to people who can operate the newest tools. It belongs to educators, technologists, leaders and domain specialists who can work across boundaries while keeping learners, evidence and public responsibility at the centre.

A new graduate programme marks a consequential transition

The National Institute of Educational Planning and Administration in India scheduled the first classes of its Master of Arts in Artificial Intelligence in Education for 10 August 2026. The official notice describes a two-year, full-time programme integrating artificial intelligence with education to prepare graduates for learning and policy.

This is verified programme and calendar information from a public university-level institution. At publication time, it establishes the intended commencement date and the programme's stated purpose. It does not establish enrolment quality, curriculum effectiveness, graduate competence, employment outcomes or institutional impact.

The development is nevertheless important. AI in education is no longer only a short workshop topic attached to computer science, instructional design or educational technology. It is becoming a named field of advanced study situated within educational planning and administration.

The serious question is therefore not whether institutions need expertise. It is what kind of expertise deserves the name. A new qualification should not reproduce the fragmentation it is meant to solve.

The field must be educational before it is technological

An AI-in-education professional should understand models, data, prompting, automation and system limitations. Yet technical knowledge becomes educational only when it is connected to questions of purpose: who is learning, what capability should develop, why it matters, how evidence will be gathered and who remains responsible for the result.

Personalisation is not automatically good teaching. Faster feedback is not automatically accurate feedback. A polished generated resource is not automatically a sound curriculum. An analytics dashboard is not automatically a trustworthy account of a learner.

Educational judgement asks whether a tool strengthens attention, understanding, practice, feedback, transfer and belonging. It also asks what the tool may displace: teacher knowledge, learner agency, local language, human relationship, independent capability or a route to appeal.

The field should therefore begin with learning theory, curriculum and human development rather than treating them as soft additions to a technical core. Education supplies the ends by which the technology must be judged.

AI competence is multidimensional

A 2026 Ghanaian study surveyed 509 preservice teachers who had completed AI-related modules. It examined six domains: knowledge, pedagogy, assessment, ethics, human-centred education and professional engagement. Pedagogical and assessment indicators operated as important bridges within the competence structure, and the analysis identified more than one competence profile among participants.

The research is valuable because it refuses to reduce competence to tool use. Its results are based substantially on self-report within one university, however, and cannot establish how consistently participants would perform in classrooms or whether the reported competence would persist over time.

The six-domain structure offers a useful starting point for professional formation. Technical understanding should be tested alongside the capacity to design learning, judge evidence, recognise ethical risk, preserve human agency and work responsibly with colleagues and communities.

Institutions should expect uneven profiles. A strong technical user may need deeper curriculum knowledge; an experienced educator may need data literacy; a policy specialist may need authentic classroom practice. Professional preparation should diagnose and develop the whole profile rather than issue one undifferentiated certificate.

African capacity-building is expanding, but participation is not yet impact

Kenya's Centre for Mathematics, Science and Technology Education in Africa announced 2026 AI-literacy training for 2,509 senior-school teachers from twenty-five counties. The programme connected AI foundations, hands-on tools, learner-centred competency-based education and responsible use.

UNESCO has also reported an expanding digital- and AI-skills pathway in Kenya's vocational system, moving from educator preparation towards learner engagement. Its Global Skills Academy partnership intends to reach teachers and learners across several developing economies.

These are consequential implementation efforts and provider reports. They establish reach, design intentions and participation targets, not changed teaching, valid assessment, equitable learner outcomes or long-term institutional capability.

The next professional step is to connect training to enacted evidence. Every course should lead to a revised learning sequence, an assessed learner performance, a data-protection decision, an accessibility improvement or a documented pilot. Completion records should open an evaluation pathway rather than close the question of competence.

Interdisciplinary curriculum requires governed collaboration

Recent curriculum research is beginning to describe how such collaboration can be designed. A 2026 engineering study developed an interdisciplinary AI curriculum through curriculum mapping and expert review with educators and industry. The authors report positive formative validation while explicitly noting that implementation effectiveness still requires study.

Another governance-oriented framework for interdisciplinary curriculum design positions large language models as bounded collaborators within a human-governed process spanning public policy, technology and service design. Its contribution is architectural rather than proof of graduate outcomes.

A survey of 898 teachers supporting a proposed vertical digital curriculum found cautious optimism alongside strong demand for regulation, training and governance. The proposed framework is adaptable and promising, but it too requires validation in real settings beyond the context from which it emerged.

The shared lesson is that interdisciplinarity does not occur because several departments appear on a programme document. It requires an explicit capability map, shared problems, agreed standards, integrated assessment and legitimate academic decisions about what counts as evidence.

Curriculum and assessment must remain inseparable

A responsible programme should prepare learners both to work with AI and to evaluate AI-mediated education. They should design a learning experience, explain the theory governing it, examine the system's outputs, identify affected groups, test accessibility and present evidence about what changed.

Assessment should combine an artefact with process records, oral explanation, observed practice and independent capability. A capstone might require a bounded pilot in a real educational setting, followed by evidence of learner experience, error analysis, cost, staff workload, data handling and a continuation-or-closure recommendation.

This is especially important in competence-based education. Knowing the vocabulary of bias, prompting or personalised learning is not the same as making a defensible decision when a learner is misclassified, a model invents a source or a teacher disagrees with an automated recommendation.

The professional claim should be proportionate. A graduate may be ready to analyse, design and coordinate responsible implementation without being a software engineer, lawyer, psychologist or subject expert. Competence includes knowing when another profession must decide.

E-learning expertise must include access, continuity and exit

AI-enabled education is often presented through ideal connectivity, current devices and fluent use of dominant languages. Many learners encounter intermittent electricity, costly data, shared phones, inaccessible interfaces and systems that perform unevenly in local languages.

An AI-in-education professional should be able to design for low bandwidth, downloadable and printable materials, asynchronous participation, assistive technologies, language diversity and human support. The goal is not to imitate a high-resource classroom on a smaller screen but to build a resilient learning ecology around the realities of the learner.

Continuity also matters. Institutions need to know what happens when a vendor changes price, withdraws a product, updates a model or restricts data access. Learner records, assessed work and essential curriculum should not disappear with a subscription.

Exit planning is therefore an educational competence. A system is not sustainable merely because it can be purchased this year; the institution must be able to preserve learning, records and human capability when the tool changes or leaves.

Governance is part of professional practice

A systematic review of 42 studies on ethics and governance in generative-AI education identified recurring concerns around privacy, fairness, transparency, learner wellbeing and accountability through human oversight. These principles are well established. The implementation gap remains substantial.

Educational leaders need professionals who can turn principles into decisions: map data flows, classify risk, document purpose, specify human review, test bias, inform users, handle complaints, manage conflicts of interest and set renewal or exit conditions.

The European Union's AI Act became generally applicable on 2 August 2026, subject to its phased provisions and exceptions. Its rules matter beyond Europe wherever providers, institutions or affected people fall within scope. Legal compliance, however, is not identical to educational validity or theological faithfulness.

A lawful system may still weaken learning; an engaging system may still misuse data. Responsible professionals must help governing bodies consider law, evidence, educational purpose, institutional values and the experience of affected people together.

Theological education needs a broader hybrid competence

Theological colleges face all the educational questions already described and several more. They form people for interpretation, pastoral care, public witness and accountable ministry, not only for the production of academic text.

Recent scholarship on theological education and generative AI describes constructive possibilities in prompting, course design and institutional experimentation. Research in religious education similarly argues that AI may support engagement and access but cannot replace witness, accompaniment and participation in a faith community. These are theological and pedagogical arguments, not controlled evidence of formation outcomes.

A theologically formed AI-in-education leader should understand Scripture, doctrine, ecclesial authority, spiritual formation, safeguarding and the limits of educational assessment. The person should be able to distinguish a learning achievement from calling, character, licensing, appointment or ordination.

Competency-based theological education can provide an important discipline: connect knowing, being and doing; use authentic ministry evidence; involve mentors and communities; and refuse to treat a fluent generated artefact as proof of pastoral competence. AI may assist learning while human communities remain responsible for formation and recognition.

Digital ministry needs stewards, not synthetic authorities

Churches and ministries increasingly use AI for communication, translation, research, administration, teaching resources and online engagement. The work often falls between media teams, pastors, educators, technology volunteers and governing bodies.

A boundary-spanning steward can help define what the system may do, what data it may receive, how theological claims will be verified, when material AI assistance should be disclosed and where a human person must take over.

No educational or technical qualification authorises a person to simulate prophecy, manufacture testimony, diagnose spiritual conditions or automate consequential pastoral decisions. Digital competence must remain under spiritual, ethical and organisational accountability.

The deepest professional skill is translation: helping a pastor understand a data risk, helping a developer understand a formation purpose, helping a board understand an assessment claim and helping a learner know where the machine's role ends.

Design a role before recruiting a title

Institutions may call the role an AI-in-education lead, learning-technology steward, digital pedagogy specialist or responsible-AI coordinator. The title matters less than the mandate, authority, relationships and evidence expected from the work.

A useful mandate covers eight connected domains: educational purpose, curriculum, teaching, assessment, inclusion, technology and data, governance, and evaluation. It should identify which decisions the professional may make, recommend or escalate.

The role should not become an alternative academic senate, technology department, data-protection office or ministry authority. Its purpose is to coordinate legitimate expertise and keep implementation joined to learning. Formal committees remain responsible for the decisions entrusted to them.

Institutions also need succession and distributed capability. One enthusiastic person cannot become the only interpreter of every system, contract and pedagogical choice. Faculty, administrators, learners, technical staff and governors need appropriate competence for their own responsibilities.

Evaluate the profession through institutional contribution

The success of this emerging field should not be measured mainly by tool adoption, certificates issued or the number of generated resources. Those indicators show activity and reach, not educational value.

Evaluation should examine learner performance and independent transfer, teacher practice, assessment validity, inclusion, complaints, data incidents, staff workload, cost, vendor dependence and the quality of governing decisions. It should also document where the professional recommended that an implementation stop.

Longitudinal research is essential. Graduate programmes and institutional roles should be followed into practice to determine which competence combinations predict responsible implementation and whether their contribution survives leadership turnover, platform change and the end of external funding.

Global South institutions should lead this research. Their multilingual learning environments, mobile-first practices, communal knowledge traditions, resource constraints and governance realities are not marginal conditions to be corrected before knowledge can be produced. They are sites from which the world can learn how technology must answer to education under real constraint.

Use a boundary-spanning formation pathway

A strong formation pathway begins with educational purpose and human development. It adds technical and data literacy, curriculum and assessment design, inclusive e-learning, ethics and law, organisational change, disciplinary application and supervised practice.

Learners should work in multidisciplinary teams while remaining individually answerable for defined judgements. They should encounter failed pilots, inaccessible systems, disputed evidence and vendor claims rather than only polished demonstrations.

Every programme needs an external advisory structure including educators, learners, technologists, legal and data-protection expertise, employers or ministry contexts, disability perspectives and communities affected by implementation. Participation should be transparent, and commercial interests should be disclosed.

A culminating decision should ask not merely, 'Can this be built?' but, 'Should this be used here, for these learners, under these conditions, with this evidence and this accountable human authority?' The ability to answer that question is the profession's public value.

Build the bridge without creating a new silo

The emergence of formal study in artificial intelligence and education is a welcome sign of maturity. Education systems need people who can move beyond improvised experimentation and connect technological possibility with curriculum, evidence, ethics and governance.

The field will fail if it becomes another isolated office speaking a specialist language to itself. It will flourish when its professionals are technically credible, educationally grounded, ethically serious, contextually intelligent and humble about the authority of other disciplines and communities.

Africa and the wider Global South should not wait to import a finished profession. We can define it through our own educational purposes, languages, theological traditions, public responsibilities and evaluated practice while learning critically from global standards and research.

The tool expert asks what the system can do. The educator asks what people should learn. The governor asks who is accountable. The emerging professional must help all three hear one another—and ensure that the learner is never lost between them.