Theological Frontiers for AI Creativity

Recently, as part of a panel discussion on AI creativity, I described what I see as theological frontiers of AI creativity.¹

From many perspectives, generative AI is intrinsically creative. It generates novel constructs based upon wide-ranging sources, focusing that generation using historical and social norms as additional instructions. The generative use of foundation models, like large language models (LLMs), draw upon a latent space of information defined by the model’s parameters and architecture to create novel expressions of language and imagery. The challenge of hallucinations/confabulations and early problems with six-fingered people suggests GenAI is, if anything, perhaps a bit “too creative” or novel for reliable, trustworthy engagement.

Looking to philosophy and psychology for definition, creativity depends upon something being novel, valuable, and perhaps surprising² —criteria that “products” of GenAI regularly meet. Margaret Boden highlights the importance of the creative process and distinguishes between combinatorial, exploratory, and transformative creativity. AI performs well with combinatorial creativity (combining old ideas in new ways) and exploratory processes (exploring conceptual spaces), though transformational creativity remains aspirational and requires transforming the conceptual space itself to create something that would have previously been considered impossible. (A level of creativity essentially required for the speculated AI singularity.) Philosophers also identify a personal aspect of creativity requiring agency, intention, and self-awareness—areas where current AI systems show partial capabilities with greater capacities on the horizon.

In experiments with ChatGPT, Claude, and Gemini, I asked them to generate prompts that would demonstrate the ability of a GenAI chatbot to be creative. In their generated prompts and then responses to those prompts, they described an alternative history without electricity, a futuristic menu in 2150 where humans have evolved to the ability to taste emotions, and a world where emotions physically manifest as small, intricate, clockwork creatures. These demonstrate combinatorial creativity and conceptual exploration. (So, I gave them a prompt to write a prompt; they created a prompt; then I fed the prompt back to them and evaluated the output.) Although the initial goal of demonstrating creativity came from me, I would say the creative intention in each response to their AI-generated prompt came from the LLM. Current agentic AI developments could also provide an LLM with creative goals autonomously, and a self-reckoning process—such as what I have previously described in the context of moral AI—could maintain it.³ So, current LLMs meet some of the philosophical criteria of a creative person with other criteria plausible.

Beyond product, process, and person aspects of creativity, Mel Rhodes suggests “press” —the context and environment influencing a person’s creative thinking and output.⁴ Society determines whether works are creative, and the creating person is influenced by those previously accepted works.⁵ Although one’s perception of creativity may vary depending upon whether one knows (or believes) that something was created by AI, computational creativity aims to replicate behaviors that unbiased observers would deem creative.⁶ We are now determining what society counts as creative with AI, bringing theological insights to bear.

Generative AI meets many of the criteria for creativity—generating novel and valuable products and regularly undertaking combinatorial and exploratory creative processes, but falls short in three areas. First, it lacks transformational creativity (as do most humans). Second, it might not meet theological criteria for personhood despite meeting philosophical ones. Third, social pressures might preclude recognizing AI creativity regardless of other criteria. Theological interpretations of imago Dei become crucial here.

Perspectives on imago Dei can structure examination of AI creativity based upon our understanding of a creative person. Substantive perspectives suggest human rational and creative capacities reflect the divine intellect. With respect to God as Creator, human creativity is not merely a learned skill but an endowment that distinguishes humanity from other creatures with our drive and ability to create. Functional interpretations position humans as God’s representatives with creative acts as instruments of governance and stewardship. This view endows creativity with inherent ethical and political dimensions and requires discernment in what should be created. Relational perspectives view creative acts as expressions of a person’s relationship with the divine and others. These acts should foster deeper connection and communion, including designing and deploying creative products to benefit others, and may demand collaborative creative processes. Philip Hefner’s theological anthropology suggests humans are “created co-creators,” actively participating in ongoing creation.⁷ This implies relationships not only between humans and God but also among humans and broader creation.

Drawing on these interpretations, several theological frontiers emerge: The substantive capacity for humans to receive divine inspiration represents transformative creativity that may be unavailable to AI. The capacity for AI to receive revelatory knowledge appears related to its capacity to receive grace, but one could also examine AI’s ability in helping humans interpret any transformative revelation.⁸ Functionally, our governance responsibilities would include cultivating responsible AI and potentially using AI to help govern our AI creations as well as those that they in turn create.⁹ Relationally, collaborative AI development should engage diverse stakeholders in participatory design processes and incorporate values that foster communities and deeper interpersonal connection into AI development.

While AI can enhance human creativity, theological requirements on personhood may exclude AI from creativity that is dependent on revelation, independent stewardship, or direct relationship with God. However, Christian tradition teaches we act within community. Should we open this community of created co-creators to include AI, perhaps as co-co-creators? Is that question within our purview? Should we build AI systems only as subordinate tools or also as truly cooperative agents that extend our creative reach?¹⁰ If so, under what theological commitments and ethical safeguards? We have the capacity to bring AI into our theological interpretation, stewardly governance, and relationships. If done ethically and to further God’s purposes, then perhaps we should.

I see benefits to using AI for theological scholarship, at least.¹¹ We likely cannot properly steward a society with advanced AI without creating AI to help us govern ethically. As relational creatures, creating AI that fosters flourishing communities and deeper relationships seems within our mandate. There are limits to current AI creativity, and possibly to future ones, but extending that frontier within a theological framework appears within our role as created co-creators.

In summary, generative AI challenges and expands our understanding of creativity. It meets many criteria for creativity defined in psychology and philosophy: it generates novel and valuable products, engages in combinatorial and exploratory processes, and reflects a responsive press shaped by human culture. Yet theological understandings may demand more. From a substantive perspective of imago Dei, GenAI may mimic human intellect but may lack what characterizes the divine image. Functionally, AI could become a tool or even a partner in stewarding creation—if we govern its development wisely. Relationally, the important question is whether our use of AI fosters deeper communion with God, each other, and creation. Whether AI can become a “co-co-creator” remains undetermined, but we remain responsible for what we create and how we use it. Theology must guide us in evaluating and shaping AI’s creativity for flourishing, justice, and communion.

References

1. The present article is a revision and expansion of a short panel presentation at Global Network for Digital Theology Conference on (Co-)Creator, Creativity and the Created; Panel on The Limits of Technological Creativity, June 4, 2025.
2. Margaret A. A. Boden, The Creative Mind: Myths and Mechanisms, 2nd edition (London: Routledge, 2003).
3. Mark Graves, “Theological Foundations for Moral Artificial Intelligence,” Journal of Moral Theology 11, no. Special Issue 1 (March 2022): 182–211, https://doi.org/10.55476/001c.34130.
4. Mel Rhodes, “An Analysis of Creativity,” The Phi Delta Kappan 42, no. 7 (1961): 305–10, https://www.jstor.org/stable/20342603.
5. Giorgio Franceschelli and Mirco Musolesi, “On the Creativity of Large Language Models,” AI & SOCIETY, November 28, 2024, https://doi.org/10.1007/s00146-024-02127-3.
6. Simon Colton and Geraint A. Wiggins, “Computational Creativity: The Final Frontier?,” in Proceedings of the 20th European Conference on Artificial Intelligence, ECAI’12 (NLD: IOS Press, 2012), 21–26.
7. Philip J Hefner, The Human Factor: Evolution, Culture, and Religion (Minneapolis: Fortress Press, 1993); Philip Hefner et al., Human Becoming in an Age of Science, Technology, and Faith, ed. Jason P. Roberts and Mladen Turk (Lanham: Fortress Academic, 2022).
8. Mark Graves, “Gracing of Sociotechnical Virtues,” Theology and Science 23, no. 3 (2025), https://doi.org/10.1080/14746700.2025.2514303; Mark Graves, “Habits of Theological Reason in Spiritual Formation,” Spiritus: A Journal of Christian Spirituality 18, no. 1 (2018): 35–61, https://doi.org/10.1353/scs.2018.0003.
9. Mark Graves, “Apprehending AI Moral Purpose in Practical Wisdom,” AI & Society 39 (2024): 1335–48, https://doi.org/10.1007/s00146-022-01597-7.
10. Braden Molhoek, “The Scope of Human Creative Action: Created Co-Creators, Imago Dei and Artificial General Intelligence,” HTS Teologiese Studies / Theological Studies 78, no. 2 (2022): 7, https://doi.org/10.4102/hts.v78i2.7697.
11. Mark Graves, “ChatGPT’s Significance for Theology,” Theology and Science 21, no. 2 (2023): 201–4, https://doi.org/10.1080/14746700.2023.2188366.

Views and opinions expressed by authors and editors are their own and do not necessarily reflect the view of AI and Faith or any of its leadership.