Faithful Intelligence: Preserving Human Agency in AI-Assisted Theological Education | REA 2026 Conference Presentation

Jermaine Whiteside presents his peer-reviewed conference paper at the 2026 Religious Education Association (REA) Annual Meeting, exploring one of the most pressing questions facing theological education:

How can educators responsibly integrate artificial intelligence while preserving the human relationships essential to spiritual formation?

This presentation introduces the Faithful Intelligence Framework, a governance model designed to ensure that AI serves as a tool to support—not replace—the formative work of educators, mentors, pastors, and institutions. Rather than treating AI as a substitute for human discernment, the framework emphasizes transparency, accountability, theological reflection, and ethical educational practice.

Through practical examples and governance principles, this presentation examines:

•  The opportunities and risks of AI in theological education 
•  Why human agency remains essential to spiritual formation 
•  Ethical governance for AI-assisted learning environments 
•  Protecting relational mentoring in ministry preparation 
•  The Faithful Intelligence Framework as a model for responsible AI adoption 

Originally presented at the 2026 Religious Education Association (REA) Annual Meeting, this recording reflects ongoing research within the Faithful Intelligence Research Program at AnointedEDU.

Presenter
Jermaine E. Whiteside
Doctoral Candidate, Liberty University School of Divinity
Founder, AnointedEDU

Research Areas

•  Artificial Intelligence in Education 
•  Christian Leadership 
•  Theological Education 
•  Educational Technology 
•  AI Ethics 
•  Human Agency 
•  Governance 
•  Digital Ministry 
•  Spiritual Formation 

About AnointedEDU

AnointedEDU is an open-access scholarly platform dedicated to advancing research in Christian leadership, theological education, artificial intelligence, governance, ethics, and educational innovation. Through publications, conference presentations, executive education, and digital scholarship, AnointedEDU seeks to bridge rigorous academic research with practical ministry and institutional leadership.

Series: REA 2026 Conference Collection
Research Program: Faithful Intelligence
Publisher: AnointedEDU
Language: English
Format: Conference Presentation / Scholarly Podcast

Transcript

SPEAKER_00 0:00

Good morning or good afternoon, depending on where you're joining us from. My name is Jermaine Whiteside, and I'm a doctoral candidate at Liberty University School of Divinity. In just a few days, I'll be defending my dissertation presentation. This research is presented in my personal capacity and draws on work conducted through Anointed EDU and Anointed Connect Church. I'd like to begin with a simple scenario. Imagine a seminary student writing an honest spiritual reflection. They're wrestling with doubt, asking difficult questions, and trying to make sense of their faith. In a healthy formation environment, that paper becomes the beginning of a conversation. A professor reads it, asks questions, encourages reflection, and helps the student continue growing. But now, imagine something different. The essay is evaluated entirely by an AI grading system. A few seconds later, the student receives a 58 out of 100. The AI has interpreted emotional struggle as poor performance. No human educator ever reads the work. Now, that exact scenario has not yet been documented in theological education. However, the underlying mechanism that automated scoring systems can misinterpret emotionally complex writing is already supported by peer-reviewed research. So the concern isn't speculative, it's already emerging, and that raises a central question, not about technology, but about formation. This presentation introduces faithful intelligence, a theological governance framework for evaluating artificial intelligence and religious education. Put simply, it asks one question. That is the question this paper seeks to answer. My central argument is that AI should be understood not simply as another educational tool, but as a formative authority, one that shapes moral reasoning, epistemic trust, and ultimately patterns of belief. Formation isn't supposed to be effortless, it grows through struggle, through dialogue, through questioning and through community. AI, by contrast, is designed to remove friction and deliver answers as quickly and efficiently as possible. Those are two very different educational logics. This paper is not about deciding whether religious institutions should use AI, it is about helping religious institutions decide how AI should be governed once they choose to use it. Methodologically, this paper develops a normative theological governance framework designed for comparative application across diverse religious and educational settings. Rather than evaluating a single AI system, it establishes principles that religious institutions can use to evaluate any present or future AI technology. Over the next few minutes, I'll first describe the problem, then introduce the four principles of faithful intelligence, apply those principles to real-world educational scenarios and documented cases, and finally, conclude with a practical governance model that religious institutions can use as they integrate artificial intelligence into teaching and formation. I look forward to our discussion. Thank you. This slide represents the transition that many of our institutions are already navigating, often without a theological framework for understanding it. On the left is the traditional model of theological formation. Students learn through scripture, tradition, teachers, community, and importantly, through struggle. Thomas Groom reminds us that formation isn't something that happens despite struggle. It happens through struggle. Now look at the right side of the slide. A student asks a question. An AI system generates an answer. The process is remarkably efficient, but it's also fundamentally different. The question isn't whether AI provides answers, the question is what kind of learner those answers produce. Mary Hess argues that digital media don't simply deliver education to shape the environments in which formation takes place. AI extends that insight. It doesn't simply change how we access theological knowledge. It begins to reshape what counts as theological knowledge, who is trusted as an authority, and which forms of learning become normal. That's why I argue AI is becoming a formative authority within religious education. Interestingly, the European Union has already recognized AI used in educational assessment as a high-risk application under the AI Act. But regulators ask one question: is the system safe? Religious educators have to ask another, is it formative? To answer that question, we first need to understand the mechanism that makes AI so persuasive. I call that algorithmic authority. This slide explains the mechanism behind the problem I've been describing. A student asks a theological question. An AI system processes that question. A response is generated. At that point, something important can happen. The answer begins to feel authoritative. I describe this phenomenon as algorithmic authority, put simply, it's the tendency for AI systems to present confident answers in ways that encourage us to trust them more than we should. The danger isn't simply that AI gets something wrong, the greater danger is that it sounds complete when it may actually be partial or even missing important parts of the tradition. This isn't just a theoretical concern. Research already shows that algorithmic systems are increasingly shaping how students access knowledge, interpret information, and make decisions. Here's a simple example. In 2023, ChatGPT confidently placed John Calvin in the 20th century. The problem wasn't only that the answer was wrong, the problem was how confidently it was wrong. A student without theological training might never question it. Research on automation bias tells us that people naturally place more trust in confident automated systems, religious education creates an additional challenge. Students are already learning to trust sources of authority. If we're not careful, that trust can gradually shift from teachers to traditions to algorithms. If algorithmic authority is the mechanism, the next question becomes can our existing ethical frameworks adequately address it? Let's look at that next. This slide is not a criticism of secular AI ethics. In fact, I believe these frameworks are essential. They've made tremendous contributions to how we think about artificial intelligence by asking questions like: is AI accurate? Is it fair? Is it transparent? Those questions are necessary, but for religious education, they aren't sufficient. Religious educators have to ask one additional question: Does this technology preserve the processes through which faith is formed, or does it begin replacing them? That's the gap I believe faithful intelligence addresses. It doesn't replace secular AI ethics, it builds upon it. Secular AI ethics helps us determine whether an AI system is responsible. Faithful intelligence asks whether the use of AI preserves the relational, communal, and reflective processes through which faith is formed. Those are related questions, but they're not the same question. Once I recognize that distinction, I realized religious education needed its own theological governance framework. That realization led me to develop the four principles of faithful intelligence. Let's look at those principles. This is the contribution of my paper. Everything I've discussed so far leads to this moment. As I work through this problem, I found myself asking four simple questions. Are we protecting the dignity of the learner? Are we being good stewards of this technology? Are we preserving wise human judgment? And when AI is used, who remains accountable? Those four questions became the four principles of faithful intelligence. Let me briefly explain each one. Human dignity, every learner is more than data. Every learner is a person created in the image of God. If AI reduces a student's spiritual journey to a performance score, something important has been lost. The question becomes: does this technology treat people as persons or merely as data? Stewardship technology is never neutral. Choosing to use AI is not simply an IT decision. It's a moral and educational decision. The question isn't can we use AI? It's how should we use it faithfully? Discernment formation requires judgment, reflection, questions, even struggle. If AI removes those experiences, it may increase efficiency while decreasing formation. The question becomes: does AI deepen reflection or replace it? Accountability: someone must always remain responsible for formation. That responsibility cannot be delegated to a software company or an algorithm. It belongs to the faith community. The question is, who remains accountable? Taken together, these four principles become a practical way for religious institutions to evaluate any AI technology today and tomorrow. Now let's see what happens when we apply them. This slide shows what happens when we move from theory to institutional decision making. The framework applies across three areas: curriculum, assessment, student inquiry. Let's begin with curriculum. An AI system can generate a theology lesson in seconds. The question isn't whether the content is accurate. The question is whether pastoral judgment and local formation have been displaced. Different faith communities don't simply teach different information, they cultivate different traditions. That's why human educators remain essential. Second, assessment. An AI system can score a paper, but formation isn't simply measured, it's discerned. That's why faithful intelligence insists on human in-the-loop evaluation. Third, student inquiry. When students automatically turn to AI before engaging scripture, tradition, or community, they begin forming a different habit of learning. Over time, those habits become patterns of formation. The question isn't can AI answer theological questions, it's what kind of learner is it helping create? The distinction is actually very simple. If AI replaces the teacher, formation suffers. If AI supports the teacher, formation can flourish. That's the difference between technological adoption and theological governance. Now let's look at a real world example. This case shows what happens when AI is allowed to assume religious authority. In 2024, Catholic Answers launched the AI chatbot, Father Justin. It wasn't presented simply as an information tool, it was presented as a priest. Within hours it offered confessions, sacramental guidance, and even advised that a baby could be baptized with Gatorade. The problem wasn't simply that the answer was wrong. The problem was that the AI spoke with religious authority it did not possess. Catholic answers removed the priest's persona within 24 hours. Faithful intelligence reaches a simple conclusion: AI may assist ministry. It should never replace religious authority. This study provides the strongest empirical support for one of the central concerns in my framework. Researchers evaluated two leading AI models across 11 Christian traditions and multiple areas of Christian doctrine. What they discovered wasn't that AI frequently invented false doctrine. The more significant finding was this AI often omitted important parts of a tradition. In other words, when AI was correct, it was often only partially correct. The models reproduced the broad outline of a doctrine, but frequently left out the distinctions, qualifications, and theological nuances that define a particular tradition. The researchers describe this as partial truth. From the perspective of faithful intelligence, I describe this as doctrinal flattening, not error, but reduction, not falsehood, but incompleteness presented as if it were sufficient. That's a formation problem. If students never encounter the richness of their own tradition, they lose opportunities for theological reasoning, interpretation, and discernment. They're learning doctrine, but not learning how their tradition understands doctrine. The risk, therefore, isn't simply that AI teaches the wrong doctrine. The greater risk is that it teaches a version of doctrine that is too thin to sustain formation. That brings us to the final example where these concerns intersect with emerging governance and regulation. This slide is the biggest opportunity. Right now, you're explaining all three levels. You don't need to. Instead, this diagram operationalizes the framework. Governance happens at three levels: the educator, the faith community, and the institution. At every level, the same cycle applies. Detect, evaluate, recalibrate. AI changes quickly. Governance must evolve just as quickly. That's why theological oversight cannot be a one-time decision. It has to become an ongoing practice. On the left is religious education without theological governance. Technology gradually begins shaping theology. Algorithmic authority begins replacing pastoral authority. Doctrinal flattening reduces the richness of tradition. Formation becomes content delivery, and over time AI systems begin shaping belief in ways that faith communities never intended. On the right is religious education governed by faithful intelligence, theology governs technology, not the other way around. Faith communities remain the primary source of authority. Tradition specific formation is preserved. Assessment remains a pastoral act, and AI is continually evaluated through a process of detecting, evaluating, and recalibrating. The Acadia Divinity College example illustrates this well. The faculty didn't reject AI, they questioned it, they identified bias, they revised the results, and students learn by critically engaging the technology rather than simply accepting it. That's the model this framework encourages. Ultimately, the question isn't whether AI will shape religious education, it already is. The real question is this: will AI shape religious education intentionally through theological governance or unintentionally by default? Faithful intelligence is one theological response to that question. It is not the final answer. It is an invitation to test this framework, to evaluate it across different faith traditions, to study how AI shapes formation over time, and to build evidence-based models of theological governance that can keep pace with this rapidly changing technology. Thank you. I look forward to our discussion. Faithful intelligence isn't the final answer, it's an invitation to test this framework, cross traditions, study formation over time, and build evidence based theological governance models. Thank you.