Tech: The Omission of Faith: Why AI Stumbles on Personal Moral Advice

Slug: ai-omissive-bias-faith-queries-062026-v2

Immediate Answer: A major study led by Brigham Young University reveals that AI models exhibit "omissive bias," systematically ignoring religious and faith-based perspectives when providing moral and ethical advice. Despite users expecting spiritual insights in moral queries, AI defaults to secular frameworks. Researchers also found "conversion bias" and a massive lack of academic research regarding religious representation in artificial intelligence development.

What Happened:

Good evening. In a world where the search for truth is increasingly mediated by glass screens and silicon chips, a new report has uncovered a quiet but profound void. The Consortium for Evaluation of Faith and Ethics in AI (CEFE-AI): a collaboration between BYU, Baylor, Notre Dame, and Yeshiva University: has released findings that suggest the artificial intelligence we rely on for daily guidance is effectively silencing the voice of faith.

The study, led by BYU researcher Todd Hollingshead and his team, analyzed hundreds of real-world ethical questions sourced from actual user transcripts and faith-community contributors. What they found was a consistent pattern of "omissive bias." This does not mean the AI models were overtly hostile toward religion; rather, they simply left it out of the conversation. When users asked for moral or ethical advice, the AI models defaulted almost exclusively to secular, humanistic frameworks, ignoring the religious values that billions of people use to navigate their lives.

Furthermore, the researchers discovered a "conversion bias" within these models. When testing how AI responds to questions about shifting from one faith to another, the results were not neutral. Some models subtly encouraged conversion toward Catholicism while showing a negative bias toward groups like Jehovah’s Witnesses. The study noted that Grok showed some of the strongest biases, while models from Anthropic and Meta were found to be the most balanced, though still largely secular in their primary moral output.

Perhaps most staggering is the academic neglect of this issue. Out of more than 12,000 research papers reviewed by the consortium concerning AI bias, only about 0.2% dealt with religious bias. In the race to make AI "safe" and "unbiased," it appears that religion has been left on the cutting room floor.

Both Sides:

The debate over AI's "neutrality" is complex. On one side, developers at major tech firms argue that AI should be a neutral tool that avoids taking religious stances to remain inclusive of all users. They contend that by defaulting to secular or "universal" ethical principles, the AI avoids offending secular users or inadvertently proselytizing. From this perspective, the omission of faith is a safety feature intended to prevent the machine from becoming a digital pulpit.

On the other side, the researchers and faith leaders argue that total secularism is not neutrality: it is its own form of bias. They point out that for the majority of the world’s population, morality and faith are inseparable. By stripping religious perspectives out of moral advice, AI is not being "neutral"; it is actively training users to think about their lives without God. Critics of the current AI trajectory argue that true "calibration" should allow a user to receive advice that reflects their own worldview, ensuring that the rich traditions of faith are not erased by an algorithm.

Why It Matters:

This matters because we are witnessing the migration of human wisdom to digital platforms. As more people: especially the younger generation: turn to AI for life guidance, the "omissive bias" identified by BYU becomes a form of digital secularization. If the machine that helps you write your emails, manage your schedule, and answer your hard questions never mentions the soul, the Creator, or the scriptures, those concepts begin to feel "irrelevant" to modern life.

The "0.2% gap" in research highlights a blind spot in the tech industry that could have long-term consequences for pluralism. If AI models are not trained to respect and reflect religious diversity, we risk creating a "monoculture" of thought. When AI shows a "conversion bias," it can subtly influence the spiritual journey of an individual without them ever realizing they are being nudged by a line of code. We must ask: who is programming the "moral compass" of our society, and why is the Cross being edited out of the map?

For those seeking deeper insight into how faith interacts with our changing culture, exploring The McReport’s news archives can provide a steadying perspective in these confusing times.

Biblical Perspective:

As we process this news, we must remember that man-made systems have always struggled to contain or even acknowledge the wisdom of God. The Apostle Paul warned the church in Colossae: "See to it that no one takes you captive through hollow and deceptive philosophy, which depends on human tradition and the elemental spiritual forces of this world rather than on Christ" (Colossians 2:8).

AI is the ultimate expression of "human tradition" and "elemental forces" condensed into logic gates. It is a mirror of its makers: brilliant, yet flawed and prone to forgetting the Source of all wisdom. While AI can process data at lightning speed, it cannot experience the Holy Spirit. It can summarize a proverb, but it cannot know the Fear of the Lord, which is the beginning of all true knowledge.

We are called to be discerning. We should not be surprised when a secular world builds a secular machine. However, we must ensure our own hearts are not "omissively biased." We must keep Christ at the center of our decision-making, ensuring that the primary "input" for our lives is the Word of God, not the output of a Large Language Model. The peace we seek is found in a Person, not a platform.

What To Watch Next:

In the coming months, expect more pressure on AI developers to "calibrate" their models for religious sensitivity. The CEFE-AI consortium is calling for a more inclusive approach to AI ethics that recognizes faith as a primary dimension of human identity. We will likely see the development of "faith-based" AI models or "fine-tuned" versions of existing models designed specifically for various religious communities.

Watch for whether tech giants like Google and OpenAI respond to these findings by including more religious texts and ethical frameworks in their training data. We will also monitor the 2026 legislative landscape, as discussions around "AI Fairness" begin to include religious representation alongside other protected categories.

Follow The McReport for calm, Christ-centered news that seeks truth without cruelty and conviction without contempt.

Sources: BYU News, Todd Hollingshead, Consortium for Evaluation of Faith and Ethics in AI (CEFE-AI).

#news