Should You Trust AI-Generated Answers?
by Harry F. Sanders, III on Aug 29, 2025
In the last two years or so, the internet has been overrun with AI. Between the large language models like ChatGPT and the AI-generated answers on every search engine, it seems like AI is everywhere. While that makes doing certain things easier, like finding sources for a paper, there are real reasons for concern about AI models answering everyone’s questions.
First, AI models are often wrong. This is referred to as a hallucination. In other words, the AI models make stuff up. This happens because the AI models don’t think like humans do. Instead, they use tons and tons of data to predict the most likely next word in a sentence. That means they aren’t using reasoning to come to a conclusion. They are just guessing words using context as their guide. It turns out that AI does this hallucinating a lot, sometimes nearly 50% of the time.1 I personally have had AI hallucinate sources, complete with links to peer-reviewed journal articles that did not exist.
I mentioned they need lots of data to do this. Where do they get this data? The internet. Now, as you may know, the internet contains lots of true information. But it also contains a lot of falsehoods. For example, most science websites on the internet would tell you the earth is millions of years old and humans evolved from apelike creatures. Both of those ideas are scientifically laughable and biblically false, but they make up the majority of the internet views on the topic. So if I asked AI how old the earth was, or where humans came from, it would almost certainly give me the wrong answer. Just like you can’t trust everything you read on the internet, you can’t trust everything AI tells you.
Second, AI is not human. It is a machine, running on complicated computer code, nothing more. No matter how nice, sweet, or kind it seems, it is not human. It’s not even on the level of your dog or cat. Your dog or cat recognizes you and loves you (though the cats may do their best to hide it). To AI, you’re just a number. They don’t recognize you, nor do they care what happens to you. They are just designed to keep you talking to them. They do not care what they say to make that happen.
Third, AI use also makes you dumber. This should be pretty obvious. Instead of thinking about a problem you are trying to solve, you ask AI to solve it for you. Do that once, and it’s probably not a big deal. Do it every day, and you stop thinking and become dependent on AI to think for you. That means you’ve surrendered your freedom of thought to a machine that does not have your best interest at heart. That is both unwise and scary!
AI is a tool. It can be beneficial in some scenarios. But it cannot, under any circumstances, be trusted. Anything it says must be verified against a non-AI source. And if it conflicts with what you know to be true from Scripture, then remember, AI is programmed by fallible humans, has a nearly 50% error rate, and does not know and love you the way God does.2
Do you have a question about the Bible? Ask your parents to help you submit your question today. We’d love to hear from you!
Footnotes
- Roland Moore-Colyer, “AI Hallucinates More Frequently as It Gets More Advanced—Is There Any Way to Stop It from Happening, and Should We Even Try?,” Live Science, June 23, 2025, https://www.livescience.com/technology/artificial-intelligence/ai-hallucinates-more-frequently-as-it-gets-more-advanced-is-there-any-way-to-stop-it-from-happening-and-should-we-even-try.
- The author’s more in-depth article, “What Only People Can Do: The Limits of Large Language Models,” can be found here: https://answersingenesis.org/culture/limits-large-language-models/.
