AI is here and it is here to stay. Even for those who want to avoid the technology, it will prove increasingly difficult. Google, the most popular search engine, now gives its AI results first. Generally, these are handy and helpful summaries of search results. Of all the emerging uses for AI, using it to search the internet is likely one of the most popular ways the average user will interact with it.
Like anything Christians use as a tool, it is important to exercise discernment. Summarizing large amounts of data is the strength of machine learning and LLMs (large language models)—the technology that undergirds what is popularly called AI. As AI generated search results grow in popularity, it is especially important to understand how different AI models present biased results, especially as it pertains to questions about God.
Whether we like it or not, more and more people turn to the internet to answer their theological questions. Just like there is human tweaking that goes into how Google ranks pages that show up in its search results, there is also a fair bit of human tweaking that goes into different LLMs or AI chatbots when they give their answers.
Consider a user asking questions of an AI model about God’s existence. The answers vary widely. The answer a user gets is both based on the bias of the user and the bias of the AI model itself.
The Framing and Intent of the Asker
It is important to understand that AIs are typically tuned to tell you what you want to hear. It this way, they are biased flatterers. If a user is asking from a place of personal faith, grief, wonder, or needing hope, they can expect to get a response that supports their own faith and gives a more positive answer concerning God’s existence and care.
In the same way, if the user is coming from a combative or agnostic position, the response will focus on evolution, Big Bang cosmology, and other “established” dogmas of our secular institutions.
Calvin Smith in his series of videos called “Talk with Grok” demonstrates that framing the exploration of faith, science, and creation gives answers consistent with young earth creationist views.
Therefore, how you frame the question can determine the answer you receive.
Made in the Likeness of Its Maker
The second factor, and perhaps the largest factor, in the kind of answer you’ll receive from an AI model is how it is fine-tuned or engineered to give an answer consistent with its designer.
Anthropic by Claude has very strong guardrails against ever appearing to endorse theism. ChatGPT by OpenAI presents atheism/materialism as the “scientific” default while treating theism as a respectable personal viewpoint. Google’s Gemini appears to be the most restrictive. It has been documented as refusing to say “God exists” even in historical contexts. For example, “Did Isaac Newton believe God exists?” has triggered word-salad hedging. Grok by xAI is intentionally less censored and finds the arguments for God’s existence more persuasive than not—admitting that the evidence points to God’s existence.
What this demonstrates is that each AI answers these questions in the same way as their designers, whether that’s Sam Altman or Elon Musk. As most of these models are engineered and fine-tuned by secular agnostics or atheists, that’s what the model will present to you.
Learning Lessons
So what should we learn from this?
First, be discerning. We should remember that the internet is a mess of information from a myriad of sources. AI’s ability to summarize data is only as good as the data it summarizes, and it is also summarized according to the bias of its engineers. Results may vary. Neither the internet nor an AI response should be trusted or accepted without critical discernment. There is a wealth of knowledge at our fingertips, but it is mixed in with a good dose of lies and error. Asking the right questions is key to getting a useful response.
Second, prioritize Scripture. Scripture is the norming norm that norms all other norms. You can’t be discerning unless you know the Bible and fear the Lord (Proverbs 1:7; 9:10). Christians must remember that Scripture is our ultimate authority and that it is truthful in all that it says. The principles and precepts of God’s word are so old they are new. In this secular age, we are increasingly living in world where right is wrong and wrong is right. Therefore it is incumbent upon Christians to be Bible experts to know the way of truth in the cloud of darkness and exalt the Lord of Life as the one who brings order to chaos.
Third, speak Scripture. In the age of the machine, real answers from real people will be unique and attractive. We will be tempted to give answers that people want to hear—just like AI is programmed to do. But rather than scratch “itching ears” and add to the cacophony of “teachers to suit their own passions” we must be ready to speak the living word “in season and out of season; reprove, rebuke, and exhort, with complete patience and teaching” (2 Timothy 4:2–3). When we do this, we are giving answers consistent with the mind of our Maker, the Maker of heaven and earth and all that is in them, giving people the hope and truth of God’s inerrant word. In this way, we shine as lights in a crooked and perverse generation (Philippians 2:15).
—Tim Stephens
P.S. For more information on AI from a Christian perspective, I’ve given two lectures here and here and have written this explanatory article.