Societies are in a perpetual state of flux, undergoing continuous transformation that often makes it challenging to discern overarching shifts while immersed in the practicalities of daily life. The recent surge in the popularity and capabilities of Large Language Models (LLMs) serves as a prime example. In the immediate term, these technologies present pressing sociological questions regarding their integration into societal norms, such as the ethical considerations of using LLMs for creative or academic assistance, or the profound implications for employment and the nature of work as new forms of automation emerge.
However, beyond these immediate concerns, it is crucial to adopt a broader perspective and examine how rapid technological advancements might influence the foundational pillars of our social relationships and core belief systems. A recent collection of studies published in the prestigious journal PNAS offers a fascinating and potentially transformative insight: automated work and the rise of LLMs may be reshaping even the way humanity conceptualizes and engages with religion.
The Unseen Hand of Automation: A Global Decline in Faith?
The groundbreaking article, titled “Exposure to Automation Explains Religious Declines,” co-authored by Joshua Conrad Jackson, Kai Chi Yam, Pok Man Tang, Chris G. Sibley, and Adam Waytz, synthesizes findings from five distinct studies. One of the most compelling revelations stems from their analysis of longitudinal data spanning 68 countries between 2006 and 2019. This extensive dataset revealed a significant inverse correlation: nations with higher concentrations of industrial robots consistently exhibited lower proportions of individuals who reported religion as an important part of their daily lives. This finding suggests a subtle yet pervasive influence of technological integration on the fabric of societal belief systems, extending beyond mere economic or labor market disruptions.
The implications of this correlation are profound. While the study meticulously accounts for various socioeconomic factors that might influence religiosity, the persistent link between automation levels and religious adherence points towards a deeper psychological or sociological mechanism at play. Industrial robots, a visible manifestation of advanced automation, have steadily integrated into manufacturing and other sectors globally. Their increasing presence signifies a broader societal reliance on artificial intelligence and automated systems for tasks traditionally performed by human labor. This shift is not merely about job displacement; it hints at a more fundamental re-evaluation of human capabilities, purpose, and the sources of comfort and meaning that societies have historically derived from religious frameworks.
The Rise of Generative AI and its Spiritual Echoes
Perhaps the most startling findings emerged from the fifth study within the PNAS paper—an experiment designed to gauge the impact of learning about recent scientific and AI advancements. Participants exposed to information detailing the capabilities of cutting-edge LLMs, such as OpenAI’s ChatGPT, exhibited "greater reductions in religious conviction" compared to those who learned about general scientific breakthroughs. This experimental outcome suggests a unique potency in generative AI’s capacity to challenge or diminish religious belief, differentiating it from the broader impact of scientific progress.
The authors postulate that this phenomenon might be rooted in how people perceive AI. Unlike traditional scientific and technological advancements, which might offer new explanations for natural phenomena or provide practical tools, AI, particularly generative AI, appears to possess capacities that people may instinctively attribute to realms previously reserved for religious or spiritual understanding. These capacities could include the generation of complex narratives, the provision of seemingly insightful advice, the creation of art, or even the simulation of intelligent conversation. Such abilities might subtly erode the instrumental roles that religion has historically played in human societies—roles such as providing answers to existential questions, offering moral guidance, fostering community, and furnishing coping mechanisms in times of uncertainty or hardship.
A Historical Context of Technology and Belief
To fully grasp the significance of these findings, it is essential to contextualize them within the broader historical interplay between technology and belief systems. Humanity’s relationship with technology has always been intertwined with its spiritual and philosophical evolution. From the agricultural revolution, which gave rise to organized societies and complex polytheistic religions, to the scientific revolution, which challenged geocentric models and led to new interpretations of divine creation, technological shifts have consistently forced re-evaluations of fundamental beliefs.
The Industrial Revolution, beginning in the late 18th century, mechanization transformed not only production but also social structures, family life, and the very concept of time. While it did not directly dismantle religious belief, it laid the groundwork for secularization by shifting populations from agrarian communities with strong religious ties to urban centers, where traditional religious authority often weakened. The 20th century witnessed further technological accelerations—the advent of mass media, advanced communication, and computing—each subtly altering how information is disseminated, how communities are formed, and how individuals perceive their place in the universe.
The current era of artificial intelligence, particularly with the rapid development of deep learning and large language models, represents another profound inflection point. The Transformer architecture, introduced by Google in 2017, laid the groundwork for unprecedented advancements in natural language processing. This led directly to the public release of models like ChatGPT in late 2022, which quickly demonstrated an ability to generate human-like text, answer complex questions, write code, and even compose creative content. This level of autonomous, creative, and communicative intelligence, previously confined to science fiction, has now become a tangible reality, prompting a re-evaluation of what constitutes "intelligence" and "creativity"—attributes often seen as hallmarks of human uniqueness, and in some religious contexts, as divine gifts.
Unpacking the Mechanisms: Why AI, Not Just Science?
The PNAS study’s distinction between general scientific advances and LLMs is critical. Scientific progress, while often challenging dogmatic interpretations of scripture or natural law, typically operates within the realm of objective observation and empirical verification. It explains how the world works, often leaving the why to philosophy or religion. For instance, understanding evolution might alter beliefs about creation, but it doesn’t necessarily displace the need for moral guidance or spiritual comfort.

LLMs, however, present a different kind of challenge. By generating coherent narratives, offering seemingly empathetic responses, or even simulating philosophical discourse, they encroach upon domains traditionally associated with spiritual leaders, sacred texts, or contemplative practices. If an AI can provide a comforting response to grief, offer ethical frameworks, or generate inspiring prose, it might implicitly diminish the perceived unique value of human-derived or divinely inspired sources for these functions. The perceived "intelligence" and "creativity" of AI can mimic human cognitive processes in ways that other technologies, like industrial robots, do not, thereby potentially reducing the sense of human exceptionalism that often underpins religious belief.
Furthermore, the rise of AI could contribute to a sense of "technological solutionism," where complex human problems, including those related to meaning, purpose, and suffering, are increasingly viewed through a lens of potential technological resolution. If AI can solve more problems, alleviate more suffering (e.g., through mental health bots), or even simulate companionship, the perceived need for divine intervention or religious community might lessen.
Broader Societal Implications and Expert Reactions
The potential decline in religious conviction, as suggested by this research, carries significant broader societal implications. Religion has historically served as a powerful social glue, providing communal identity, moral frameworks, and mechanisms for social cohesion. If these functions diminish or are displaced, societies might face challenges in maintaining shared values, fostering collective action, and addressing existential questions.
While the PNAS study presents compelling correlations and experimental evidence, researchers are quick to emphasize the complexity of these interactions. Evan Stewart, an assistant professor of sociology at the University of Massachusetts Boston, whose observations on this study prompted the initial discussion, highlights that religion is a "socially powerful force" used by people to "accomplish things in the world and solve problems," even coping with hardships like job loss. The potential displacement of these instrumental roles by AI is therefore not merely an academic curiosity but a matter of significant societal concern.
Sociological experts generally acknowledge that secularization is a long-term trend in many industrialized nations, driven by factors such as increased education, economic security, and exposure to diverse worldviews. However, the PNAS study suggests that AI and automation might not just be contributing to this trend but introducing a novel, perhaps accelerated, dynamic.
From the perspective of religious institutions, these findings could necessitate a strategic re-evaluation. Historically, religions have adapted to scientific and technological advancements, often by reinterpreting sacred texts, emphasizing spiritual or metaphorical truths over literal ones, or focusing on aspects of faith that technology cannot replicate, such as community, ritual, and direct spiritual experience. The challenge posed by AI might compel religious leaders to further emphasize the uniquely human and interpersonal dimensions of faith, or even explore ways to integrate AI ethically as a tool for religious education or community building, rather than viewing it solely as a threat.
Academics in fields like AI ethics and philosophy are increasingly grappling with the implications of advanced AI on human values and societal structures. Dr. Kate Darling, a research specialist in robot ethics at MIT, for instance, has explored how human-robot interactions can evoke emotional responses, blurring the lines between tool and companion. While her work focuses on different aspects of human-robot relationships, it underscores the profound psychological impact these technologies can have. The PNAS study adds a new dimension to this discourse, suggesting that this impact extends to our deepest-held beliefs.
Looking Ahead: Policy, Adaptation, and Future Research
The findings from Jackson et al. (2023) open numerous avenues for future research and policy consideration. Long-term longitudinal studies will be crucial to establish causality more definitively and to track how religious belief evolves in societies with varying rates of AI adoption. Researchers will need to investigate cultural variations, as the impact of technology on religion may differ significantly across diverse religious traditions and geopolitical contexts. For instance, how might the impact of AI differ in highly secularized Western nations versus more traditionally religious societies in the Global South?
Policy discussions may need to address the broader societal implications of potentially declining religious adherence. If religion provides crucial social support, ethical guidance, and meaning-making frameworks, what alternative structures or narratives will emerge to fulfill these vital functions in an increasingly AI-driven world? Educators might consider how curricula can prepare future generations to navigate complex questions of meaning and purpose in an era where AI can simulate many human intellectual and creative functions.
In conclusion, the advent of sophisticated automation and Large Language Models represents more than just an economic or ethical challenge; it may be fundamentally altering the cognitive and emotional landscape in which individuals construct their core beliefs. The PNAS study offers compelling evidence that these technologies are not merely expanding our capabilities but could be reshaping our collective imagination about what is possible, what is meaningful, and what roles traditional institutions, including religion, play in our lives. As AI continues its inexorable advance, understanding its multifaceted impact on human society, including our deepest spiritual convictions, will be paramount for navigating the future responsibly and thoughtfully.








