“But solid food is for the mature, who by constant use have trained themselves to distinguish good from evil.”
— Hebrews 5:14
Hebrews was not talking about artificial intelligence. The writer was speaking about spiritual maturity: mature people learn discernment because their powers of judgment have been trained through use.
But that raises a very practical question for us today.
We can now hand many tasks that once required our own effort to AI: finding information, organizing viewpoints, comparing positions, analyzing arguments, and even producing a complete answer.
The question is not simply: Do I use AI?
Is AI helping me train my discernment, or is it doing the training for me?
Is AI Removing Busywork—or Training?
The simplest case is obvious: instead of thinking through a question myself, I ask AI to give me the answer.
But the issue is not that simple. Some people use AI not to stop thinking but to find counterarguments, identify weaknesses, compare interpretations, and deliberately challenge their own assumptions.
So perhaps the better question is not whether AI makes me think less, but this:
Some kinds of thinking are themselves necessary training. Could AI be saving me from the very part I need to learn to do?
Some work really is repetitive busywork: organizing large amounts of information, reformatting material, or locating known facts. Letting a tool handle those tasks may cost us very little.
But other kinds of difficulty are part of learning itself.
I try to explain an idea clearly and discover that I do not actually understand it. I compare two positions and realize that neither is as simple as I had assumed. I struggle to recall a passage and gradually build connections between texts. I spend a long time without an answer and learn to tolerate saying, “I do not know yet.”
So another question is worth asking:
Is AI removing pointless effort, or removing an opportunity to develop a capacity I need?
This is not merely a hypothetical concern. Recent research has begun to find that in some learning contexts students can perform better while AI is available, yet fail to show the same improvement when the AI is removed and they must answer on their own.
One randomized controlled study had university students use ChatGPT for learning and then gave them an unexpected test 45 days later. The ChatGPT group averaged 57.5% correct, compared with 68.5% in the traditional-learning group. The researchers suggested that unrestricted AI assistance may reduce the cognitive effort needed to build durable memory.
But this does not mean that using AI necessarily makes learning worse. Other studies suggest that when AI guides learners to think for themselves rather than simply doing the work for them, it can support learning.
So the question remains not simply whether we use AI, but:
Is it completing the learning for me, or helping me learn how to do it?
AI Can Do It—That Does Not Mean I Have Learned to Do It
AI can compare theological positions for me. It can analyze weaknesses in an argument. It can organize a well-structured essay. It can even take my scattered ideas and turn them into a system that sounds impressively coherent.
The finished work may genuinely be much better than what I could have produced before.
A better product does not necessarily mean a better thinker.
This is not an argument for rejecting tools. Human beings have always used books, dictionaries, commentaries, teachers, search engines, and notes to extend our abilities.
The real question is not whether I may rely on tools, but:
If the AI assistance disappeared, would I now be better than before at understanding the problem, weighing evidence, spotting weaknesses, and reaching my own conclusion?
If the answer is no, perhaps I have become capable of producing more without actually growing very much.
Spider-Man’s Suit
This reminds me of Spider-Man: Homecoming.
After Peter Parker receives the technologically advanced Spider-Man suit made by Tony Stark, he suddenly gains functions and abilities he did not have before. The film eventually forces him to face an uncomfortable question:
Without the suit, what is left?
The analogy fits AI remarkably well. My writing suddenly improves. I find information faster. My arguments become more complete. When someone asks a question, I can quickly produce a structured response.
Those improvements may all be real.
A more powerful tool does not mean that I have become equally powerful.
The subtler danger is that I may borrow AI’s cognitive power and then credit myself with the feeling of competence produced by the result.
A question that once took days to organize can now become a coherent analysis in an hour. Questions I once hesitated to answer can suddenly seem easy to discuss. Over time I may begin to think: I am good at analysis. I know a lot. My judgment is quite reliable.
But I need to ask:
How much of this ability has actually become mine, and how much is merely available for me to call upon?
Do Not Lose the Ability to Think About Your Thinking
Human beings can do more than think. We can step back and examine our own thinking.
One important feature of philosophy is that we do not only think about the world; we also question our own reasoning. Why do I think this? What grounds do I have for believing it is true? What assumptions am I making? What standards am I using to judge the evidence?
The more AI can organize questions, supply information, connect arguments, and formulate answers for us, the more important it becomes to preserve this ability.
An answer may be impressively complete, but I still need to ask:
Why am I accepting this answer?
What assumptions am I failing to notice?
If my starting point is wrong, how much does the elegant argument that follows really prove?
Am I looking for truth, or for reasons strong enough to support what I already want to believe?
We should not let having a tool that thinks extremely well slowly cost us the ability to examine how we ourselves are thinking.
Do Not Let AI Help Me Rationalize Myself
I am actually more concerned about something subtler than AI giving me an obviously wrong answer: AI can take an incomplete idea—with real blind spots—and make it extremely coherent.
I begin with my interpretation. AI supplies supporting material. I ask it to organize the argument, and it connects the pieces. I ask for objections, and it gives them. Then I ask it to answer those objections.
Eventually my original view becomes more complete, more systematic, and harder to dislodge.
Coherence is not the same as truth.
A logically organized argument proves only that the pieces fit together; it does not prove that its starting point is correct.
AI can help me think something through. It can also help me make my own position sound convincing to myself. Those are not the same thing.
Do not let AI become a tool for making my own story airtight.
Someone may respond, “Then I will simply tell AI not to agree with me. I will ask it to challenge me.” That certainly adds another layer of checking. But one limitation remains:
The AI that challenges me is still being directed by me.
I decide what question to ask, what background to provide, which line of inquiry to pursue, and when to stop.
So I still need to ask:
Am I really allowing AI to challenge my thinking, or am I gradually shaping AI into a conversation partner increasingly suited to me?
The Real Test May Be When a Person Disagrees with Me
Perhaps what exposes the problem most clearly is not how I question AI, but this:
When a real person gives me a different—even completely opposite—view, will I still genuinely listen to the reasons?
Proverbs says:
“Whoever loves discipline loves knowledge, but whoever hates correction is stupid.”
— Proverbs 12:1
And again:
“Plans fail for lack of counsel, but with many advisers they succeed.”
— Proverbs 15:22
This does not mean that someone who disagrees with me must be right, or that truth is determined by majority vote. It means that wise people know they need other perspectives to enter their thinking, and that they must remain open to correction.
At this point, the phrase “I asked AI” becomes an interesting test.
If I mean, “AI found some material; let’s examine it together,” then AI opens a conversation.
But if I mean, “I already asked AI, so this has basically been settled,” AI can become a trump card for ending the conversation.
We can even create a loop:
My understanding → I question AI from within my understanding → AI answers within that frame → I treat AI’s answer as evidence that my understanding was correct
But:
AI agreeing with me is not independent confirmation that I am right.
I may even tell someone who disagrees, “You have not studied this as deeply as I have.” “I have already looked at those sources.” “I even asked AI to analyze the opposing side.” Every sentence may be true. The other person may in fact be wrong.
But Proverbs presses a different question: Am I still a person who can be corrected, advised, and changed?
Do I Love Knowledge—or Being Right?
This may be one of AI’s subtler temptations—not because AI inevitably makes us proud, but because it suddenly makes many difficult things easy.
Answers become more complete, analysis faster, writing more polished, arguments more impressive. We can gradually acquire the feeling that we ourselves have become exceptionally capable.
Proverbs 12:1 becomes pointed here: whoever loves discipline loves knowledge.
Someone who genuinely loves knowledge can at least ask, when corrected, “Has this person seen something I have not?”
If what I really love is being right, then disagreement immediately becomes something to defeat.
So the issue is not simply how much I know, but:
Do I love knowledge and truth—or am I increasingly in love with being right?
This is also why having AI should not mean losing people.
Not because people are necessarily more accurate than AI. People also have biases, misunderstand things, and become stubborn. But a real person has one important difference:
I did not configure them.
They can reject my framing altogether and say, “I think the real issue is not even the question you are asking.”
Sometimes that is uncomfortable. But the discomfort may be exactly what I need to face.
From the emotional and relational side of the same issue, see “What’s Wrong with Talking to an AI Chatbot About Your Feelings? — Do I Only Want to Be Understood, or Am I Willing to Be Broadened?”
Not Seeing That I Am Wrong Does Not Mean I Am Not Wrong
I can think carefully, ask AI to challenge me, read scholars who disagree, and seek other opinions. After all that, I may still be unable to find the flaw in my position.
But that still does not allow me to conclude: therefore I must be right.
In 1 Corinthians 4:3–5 Paul is speaking about servants of Christ, stewardship, and who has the final authority to judge. One line is especially worth remembering:
“My conscience is clear, but that does not make me innocent. It is the Lord who judges me.”
— 1 Corinthians 4:4
This passage is not about AI, nor is Paul teaching a modern theory of critical thinking. But the posture matters:
Not seeing that I am wrong does not mean I am not wrong.
I have thought about it. I have researched it. I asked AI. I asked AI to argue against me. I asked other people too. All of those can be worthwhile.
But none of them allows me to say: therefore I must be right.
The final judgment is still not mine.
What Kind of Person Is AI Training Me to Become?
Return to Hebrews 5:14. Mature people learn discernment because their powers of judgment have been trained through use.
So after using AI for some time, I want to ask myself two questions:
Am I now better than before at understanding problems, weighing evidence, spotting weaknesses, and reaching my own conclusions?
Do I still know that I may be wrong, and do I still allow Scripture, evidence, and people around me to correct me?
Hebrews reminds us: discernment requires training.
Proverbs reminds us: wise people accept correction and know that they need counsel.
First Corinthians reminds us: even trained judgment is not the final judgment.
So if AI helps me know more but makes me need other people less; if it makes my thinking more coherent but less open to other voices; if it strengthens my arguments but turns “I asked AI” into a way of ending conversation—then I need to ask:
Has AI broadened my thinking—or merely made my world more coherent and more closed?