For me, Artificial intelligence (AI) poses challenges to Faith and Truth. AI suddenly seems everywhere—and seems posed as inevitable.
Perhaps AI is inevitable. I know in my own work, some AI can help. However, the apparent fait accompli does not obviate serious inquiry into how AI affects Faith and Truth; and, I pose, how Faith and Truth should affect AI.
I pause to note that I am not arguing against (or for) AI. Instead, I argue that, for me, caution seems prudent. I have both programmed AI systems and used AI systems. I saw first-hand the marvels (when they work) and the hair-tearing frustrations when they do not (and not always knowing why the failure).
AI, perhaps surprisingly, has been around for at least a century. However, in 2022, a new breed of AI arose using neural_networks and deep_ learning. The new systems generated, for example, large-language models (LLMs). Suddenly, ChatGPT, GROK, and other LLM-augmented AIs became household words. Suddenly, AI was seemingly and relentlessly everywhere.
About a decade before (2014), I took a course from Professor Geoffrey Hinton (deemed the godfather of AI) on something called neural networks. I programmed primitive, neural networks in that class and marveled at these strange systems and their apparently uncanny capabilities.
Yet, I most recall Dr._Hinton’s cautionary comments made then (and largely true today): we really do not know how these AI-systems do-what-they-do. In other words, science and commerce unleash something (cf. bioengineering) without perhaps knowing the consequences. We may implement and deploy AI—but we do not necessarily understand what we implement. That should cause pause as people already use these systems assuming safety, privacy, reliability, and accuracy.
Newly-elected Pope Leo XIV promptly expressed concern about AI. The Wall Street Journal recently printed a story about a physicist cautioning on un-checked AI development—and apparently supported by many high-profile individuals. An interfaith group, aiandfaith.org, discusses AI implications for some faith traditions. Companies in the AI sector hire ethics experts. Yet, cohesive analysis of the societal-effects of AI seems to lag behind both the AI implementations and the breath-taking race to AI-everything. It seems only some quiet alarms toll on the periphery.
Certainly, the obvious societal effects such as AI displacing workers or creating systems that unintentionally, or intentionally, disfavor certain outcomes remain very disturbing. Using AI to market products using AI-derived persuasion techniques threatens to drive people into more debt (a recent flap criticized an airline for alleged, AI-manipulated pricing and at least one company actively markets AI-use in sales).
AI now apparently appears in insurance, medicine, finance, and other critical systems—even driving cars.
AI-decisions already eclipse fundamental human needs. In 2024, Microsoft announced a contract to re-start the Three_Mile_Island nuclear power plant because AI requires phenomenal amounts of energy.
In 1979, I lived due-east of the TMI_reactor within the anticipated plume-area. I recall the absent classmates whose families fled and the frantic days of watching the steady-drumbeat of terrifying special-reports on TV anticipating an imminent “nuclear explosion.”
Is AI so important that we re-start nuclear reactors to power AI while placing millions at risk for centuries? Does re-naming TMI as the “Crane Clean Energy Center” pose any issues of Truthfulness?
While the societal issues disturb, perhaps more insidious is AI seemingly attacking Truth itself. For example, AI clouds the simple distinctions between doing-my-work and doing-“my”-work-“augmented”-by-AI. If AI produces outcomes that I ask for but might not even understand, is that my work? We already see this conflict unfolding with copyrights and in classrooms.
The unanticipated, and Truth-related, implications also apply to the models themselves. AI models are not “neutral” despite claims by the models’ developers. These models get “trained” on information. That information-corpus, and predicate selection thereof, shapes subsequent model output.
For example, say I train a large-language model (LLM) using the corpus of traditional Bible commentaries but omit any materials from Friends. Inquiries (prompts) to the resulting AI LLM model pull from this “learned” material to generate responses. Yet, the model outputs would likely miss the understandings of Friends. Perhaps worse, the LLM might fabricate (hallucinate) responses and attribute those fabricated responses to Friends without any qualification.
My overall point is: AI raises fundamental issues of Truth. Now augmented by AI, Truth seemingly reduces to relativistic “truth.” As AI pervades our lives in media, televisions, mobile phones, cars, electronic “books,” refrigerators, electric meters, banking, and medical offices, AI fosters dependence on the products of these systems as authoritative “truth.” (I see analogy to common quips in the 1970s, “that’s what the computer says.”) Is this reliance warranted? Can we even know if reliance is warranted?
AI continues a distressing trend, in my opinion, where Truth becomes mere many-“truths” and blithe acceptance of that relativism. When AI “needs” weigh restarting a nuclear reactor (on the site of a major nuclear accident) above the residents, I pose that there is a problem with Truth. When AI data centers pose micro-climate changes due to heating, I pose a Truth issue. If AI systems manipulate or destabilize consumer pricing or financial markets, I pose there is a problem of Truth. When we can no longer know if any photograph, any audio, any video, or any news story is authentic, I pose there is a Truth problem.
Shall we reject all AI? That may no longer be feasible without rejecting our increasingly complex society (but perhaps we need to rethink the complexity of our society). But, I think we can raise Faith-consistent voices to thoughtfully and prayerfully challenge the current “benign-paradigm” of AI and the staggering race to AI-ify everything.
Shannon Brown, Regular Attender of Seeker’s Haven WG (OYM), 1/2026
