In an era characterized by rapid technological advancements, artificial intelligence (AI), specifically publicly accessible Generative Pre-trained Transformer (GPT) AI chatbots employing Large Language Models (LLMs), raises important questions for theological scholarship. Central among these is the research question addressed by my Ph.D. dissertation: Can AI genuinely “know” God, and is theological knowledge generated by AI fundamentally different from that produced by human theologians? My thesis argues that although AI demonstrates notable strengths in generating theological knowledge – such as speed, consistency, and analytical precision – it ultimately imitates only the external form produced by human theologians and lacks the inherent spiritual dimension essential to authentic human theology.
My 9-page summary of my dissertation, which I am proposing for ETS, outlines the systematic comparative analysis done in this research. The research evaluates theological texts generated by prominent GPT-based AI chatbots – ChatGPT, Gemini, Perplexity-DeepSeek, and Claude – in dialogue with authoritative theological writings from notable theologians such as Wayne Grudem, Wolfhart Pannenberg, Veli-Matti Karkkainen, Justo Gonzalez, and Paul Tillich. This comparison is structured around several critical evaluative criteria: sources of theological knowledge, methodological approaches, justification and truth-validation processes, inherent biases, spiritual depth, and historical-contextual awareness.
My findings illustrate substantial strengths in AI-generated theological discourse, notably its impressive ability to swiftly process and synthesize vast quantities of theological and historical literature. This strength aligns closely with emerging scholarship by Mark Graves (2023), who underscores AI’s capability in enhancing accessibility and analytic scope in theology, and James Hutson (2023), who emphasizes AI’s capacity to enrich theological creativity and argument development through rapid ideation and comparative analyses.
Yet, alongside these analytical benefits, this dissertation identifies fundamental limitations inherent within AI-generated theology, echoing and extending critiques articulated by Jason Thacker (2020) and John Lennox (2020). Foremost among these limitations is AI’s inherent inability to engage in genuine spiritual experience, divine revelation, moral discernment, existential wrestling, or transformative personal reflection. As Lennox argues, AI lacks an ethical or spiritual dimension due to its intrinsically algorithmic and data-driven nature, devoid of heart, soul, and moral consciousness. Similarly, Thacker warns against an idolatrous elevation of AI, cautioning that viewing AI as a theological authority risks reducing divine truth to mere computational outputs, thereby diminishing the uniquely human relational connection to God.
This dissertation makes a distinctive contribution to contemporary theological discourse by explicitly delineating both the capabilities and profound limitations of AI in theological knowledge generation – an area still emerging within theological scholarship. By clearly articulating where AI succeeds analytically and where it fundamentally falls short spiritually, this research argues for the ongoing centrality of human theologians. Humans uniquely embody spiritual discernment, existential depth, ethical reasoning, and pastoral wisdom necessary for authentic theological inquiry. Ultimately, the dissertation advocates for a balanced integration, positioning AI responsibly as a potent research assistant but firmly subordinate to human theological authority, thus preserving theology’s inherently relational, spiritual, and profoundly human character.