Faculty Readiness for AI-Supported Teaching and Scalable Online Program Delivery in Higher Education: The EPIQ-AI Framework for Epistemic Integrity

Synthesis: Sangwa, Ndahayo, Dusengumuremyi & Mutabazi (2026) develop the EPIQ-AI Readiness Framework through an integrative secondary evidence synthesis of official statistics, large-scale faculty and institutional surveys, peer-reviewed studies, and policy frameworks published 2020–2025. The framework integrates technology acceptance (TAM/UTAUT), TPACK, and organizational readiness traditions to explain how institutions can align faculty capacity, governance, and quality assurance for AI-supported teaching and scalable online delivery. Key finding: faculty readiness is a sociotechnical alignment problem — not an individual skills deficit — and assessment, not detection, is the decisive frontier for preserving epistemic integrity.

Key Findings

1. Faculty AI adoption is widespread but shallow — 61% of faculty globally and 72% of U.S. instructors have used AI in teaching, yet only 14% of U.S. instructors report confidence using it instructionally, so adoption is not synonymous with readiness.
2. Policy maturity lags strategic ambition — only 35% of institutions have institution-wide AI policies while 40% are still discussing them, so institutional AI Governance and policy formation trail the practical presence of Generative AI in coursework.
3. Assessment is the central pressure point — detection-centered Academic Integrity regimes are unreliable, biased, and insufficient for high-stakes decisions, so validity must be rebuilt through assessment redesign toward authentic, process-rich tasks.
4. Readiness is a sociotechnical alignment problem, not an individual skills deficit — it is shaped by AI Governance, incentives, workload, AI Literacy, and course-design support rather than by "more time with the tools."

The EPIQ-AI Framework

Sangwa, Ndahayo, Dusengumuremyi & Mutabazi (2026) synthesize three theoretical traditions — technology acceptance (TAM/UTAUT), Technological Pedagogical Content Knowledge (TPACK) teacher knowledge, and organizational readiness for change — into an education-specific account of readiness centered on assessment legitimacy and epistemic integrity. EPIQ comprises four aligned readiness domains: epistemic (AI Literacy and evaluative confidence), pedagogical (Learning Design and course design support), institutional (AI Governance and Educational Development investment), and quality-and-compliance (integrity and assessment regimes). The framework posits that credible AI-supported teaching and scalable online delivery require alignment across micro-level faculty capacities and macro-level institutional systems, mediated by pedagogical and assessment design capacity.

Widespread Adoption, Limited Confidence

Evidence converges on a central pattern: readiness is uneven and adoption is not synonymous with preparedness. Globally, 61% of faculty report using AI in teaching, yet non-users cite time/resource scarcity and uncertainty about instructional application as dominant barriers (Digital Education Council, 2025). In the U.S., 72% of instructors report experimentation with GenAI, but only 14% are confident using it instructionally (Ruediger et al., 2024). Readiness is heterogeneous across faculty — shaped by prior experience and preparation rather than role or seniority alone — and TPACK self-efficacy shows a curvilinear relationship with experience, challenging the institutional assumption that "more time using the tools" closes readiness gaps (Scherer et al., 2021, 2023). Faculty are typically not stalled by ideological hostility but by an absence of time, resources, and pedagogically meaningful guidance.

Institutional Supports and Policy Maturity

Five institutional supports consistently emerge as readiness enablers: access to tools, AI Literacy training, curated best practices, clear guidelines, and an environment that tolerates failure (Digital Education Council, 2025). Yet 80% of faculty do not find institutional AI guidelines comprehensive, and only 35% of institutions have institution-wide AI policies while 40% are still discussing them (Simunich et al., 2024). Policy formation lags behind the practical presence of AI in coursework, and online expansion advances faster than capacity-building for Learning Design and learning-support roles. Governance maturity is a multi-domain institutional project spanning operations, professional development, and pedagogy. Chief online officers report faculty autonomy and buy-in as persistent barriers, meaning robust technical infrastructure will not yield scalable online quality if faculty agency is not institutionally respected and operationally supported.

Assessment, Integrity, and Governance Responses

Assessment is the decisive readiness frontier. Global faculty evidence shows 54% believe evaluation methods require significant change, and preferences tilt toward AI-permitted-with-disclosure regimes rather than blanket bans or mandatory AI use (Digital Education Council, 2025). Peer-reviewed syntheses argue higher education Assessment should shift toward self-regulated, authentic, and process-rich tasks rather than product-only evaluation (Xia et al., 2024). Crucially, detection-centered Academic Integrity regimes face empirical and ethical constraints: OpenAI discontinued its classifier after reporting false positives, detector evaluations document inconsistent performance, and GPT detectors are biased against non-native English writers, raising equity and due-process concerns (OpenAI, 2023; Elkhatat et al., 2023; Liang et al., 2023). This supports shifting away from punitive detection and surveillance tooling toward transparent policy, student AI literacy, and evidence-rich adjudication.

The Readiness Misalignment Pathway

EPIQ interprets readiness unevenness as a predictable consequence of misalignment: faculty willingness to experiment coexists with institutional underinvestment in guidelines, training, and assessment redesign, producing uncertainty and prohibition as rational defensive responses. Institutional expansion pressures, support deficits, uneven faculty response, and assessment stress combine to escalate pressure for governance redesign. Where policy clarity and training are missing, faculty may rationally restrict student AI use to preserve assessment credibility, and weak support capacity creates downstream risk for online quality assurance.

Operationalizing Readiness

EPIQ operationalizes readiness through a dashboard of threshold indicators across the four domains: structured AI Literacy pathways, applied workshops, and discipline-sensitive guidance for epistemic readiness; assessment redesign, model valid disciplinary use cases, and instructional-design partnership for pedagogical readiness; policy maturity, resourcing, workload recognition, and coherent governance for institutional readiness; and compliance with regular-and-substantive-interaction expectations plus due-process integrity workflows that privilege evidence-rich adjudication over detection alone for quality-and-compliance readiness.

What this means for practice

• Faculty developers. Build applied, discipline-sensitive workshops instead of tool access alone: 72% of U.S. instructors have experimented with GenAI, yet only 14% report confidence using it instructionally, so instructional confidence is the target rather than exposure.
• Administrators. Sequence policy clarity and faculty capacity-building ahead of AI or online scale-up: only 35% of institutions have institution-wide AI policies while 40% are still discussing them, and 80% of faculty do not find existing guidelines comprehensive.
• Instructors. Redesign assessment toward authentic, process-rich tasks rather than adopting detection tooling — 54% of faculty already see evaluation methods as needing significant change, and detectors are biased against non-native English writers.
• Administrators. Condition edtech procurement on transparent model limitations and bias-risk documentation, and fund the Educational Development and Learning Design roles that online expansion depends on.
• Faculty developers. Protect faculty autonomy and buy-in through workload recognition and shared AI Governance, since technical infrastructure alone does not produce scalable online quality.

Limitations

• The study is an integrative secondary synthesis: it generated no original dataset and involved no human participants, interviews, or surveys by the authors, so its claims rest entirely on the sources selected for inclusion.
• It draws on sources with different populations and instruments (global faculty, U.S. instructors, U.S. chief online officers), a constraint the authors state limits direct statistical comparability across estimates.
• It privileges open and authoritative sources rather than paywalled sector reports, so member-only publications limit the detail that could be extracted.
• Percentages are reproduced from cited surveys whose samples and question wording vary, so prevalence figures such as 61% and 72% are not a single measurement; the authors target robust directional patterns instead of fine-grained cross-survey comparison.

Citation

Sangwa, S., Ndahayo, C., & Dusengumuremyi, F. (2026). Faculty Readiness for AI-Supported Teaching and Scalable Online Program Delivery in Higher Education: The EPIQ-AI Framework for Epistemic Integrity. EdArXiv.