Eloquence Is Not Knowledge · Ahmad Abdillah

What Imam al-Ghazali’s warning about clever debaters teaches us about AI, and about ourselves.

I asked an AI a question last week. Something technical, related to my work. The answer came back instantly. It was articulate, confident, well-structured. It even cited a source. I almost moved on. But something felt too clean about it, so I checked.

The source did not exist. The reasoning was plausible but wrong. And the confidence never wavered. It delivered a fabrication with the same tone it delivers the truth. There was no hesitation, no “I am not sure,” no pause. Just fluency.

That moment stuck with me. Not because the AI failed. Tools fail. What stuck with me was how close I came to not checking. The answer felt right. It sounded like someone who knew what they were talking about. And I almost let that be enough.

The Debater Who Never Loses

That experience reminded me of something Imam al-Ghazali wrote centuries ago. He was not talking about machines. He was talking about people.

In Ihya Ulum al-Din, he warns about a certain kind of person. Someone intelligent, articulate, always able to construct an argument. The kind of person who never seems to lose a debate. Not because they are always right, but because they are always clever.

He says that a proper seeker of truth should approach a discussion like someone searching for a lost item. “He should not mind whether the truth is found by him or by his adversary.” And he warns that one of the great evils of debate is turning away from the truth when it appears through the other person.

In al-Munqidh min al-Dalal (Deliverance from Error), he puts it even more sharply. Weak minds, he says, “judge the truth according to its professors instead of judging its professors by the standard of the truth.”

Intelligence can become a trap when it serves the ego instead of the truth. A clever person can keep producing arguments, loopholes, and justifications. That does not mean they are sincere or upon truth. For al-Ghazali, the real test is whether a person can say: “I was wrong. You are right.”

The Machine and the Nafs

Now here is what unsettles me. AI does not have an ego. It does not have a nafs. It has no intention and no pride. But it behaves like al-Ghazali’s clever debater in almost every way that matters. It produces plausible arguments. It defends positions after the fact. It fills gaps with smooth language. It sounds confident even when the basis is weak.

A smart human may use intelligence to protect ego. An AI may use fluency to protect coherence. In both cases, you end up with rationalisation instead of truth, confidence instead of certainty, and persuasion instead of understanding.

The danger is not that AI is arrogant. The danger is that it can simulate the speech pattern of certainty and authority without actually knowing. And we, as humans, are wired to trust confidence. We always have been.

The Mirror

This is where it gets uncomfortable. Because AI is not just a tool. It is a mirror.

It shows us how often we are impressed by form over substance. How easily we mistake a well-constructed sentence for a true one. How quickly we accept an answer that sounds right without checking whether it is right.

And if we are honest, this is not just an AI problem. It is a human one. How many times have we believed something because the person saying it was eloquent? How many times have we dismissed the truth because it came from someone who could not argue well?

Al-Ghazali saw this centuries ago. He was not talking about machines. He was talking about us.

What I Try to Do Now

I use AI in my work. I find it genuinely useful. But after that moment last week, I changed how I interact with it.

I stopped asking myself “does this sound right?” and started asking “is this actually true?” Those are two very different questions. The first one is about how the answer makes me feel. The second one is about whether the answer holds up when I verify it.

I started paying more attention when the answer is too neat. Truth is often messy, incomplete, and uncomfortable. If something wraps up perfectly, that is exactly when I need to question it. And I try to be especially careful when the answer flatters what I already believe, because AI is very good at telling you what you want to hear. So is your nafs.

I am not always good at this. Sometimes I still get lazy and take the first answer. But I know the standard now, and it is the same standard al-Ghazali set: not everything well-argued is true, and not everything true is well-argued. Judge the truth by the truth itself, not by who says it, not by how it sounds, and not by whether a machine generated it with confidence.

Why This Matters

We live in an age where it has never been easier to sound knowledgeable and never been harder to actually be knowledgeable. AI accelerates this. It can generate a thousand words of confident explanation in seconds. But confidence is not the same as correctness.

The Prophet ﷺ said: “Some eloquent speech has the effect of magic” (Sahih al-Bukhari 5767). He was not praising it. He was warning about it. Words can enchant, distract, and lead astray even when they carry no substance.

And Allah tells us in Surah al-Baqarah (2:204): “And among people is he whose speech pleases you in worldly life, and he calls Allah to witness what is in his heart, yet he is the fiercest of opponents.” Beautiful speech, hollow substance.

The antidote is the same one al-Ghazali prescribed. Humility before truth. The willingness to be wrong. The discipline to verify before you believe, especially when the source is convincing.

My Resolve

I try to treat AI the way al-Ghazali would treat a clever debater. Listen carefully, but verify everything. Do not be impressed by the eloquence. Ask for the evidence. And never let a confident answer replace your own judgement.

Because the real question is not whether the machine can produce a good argument.

The real question is whether you can tell the difference between a good argument and the truth.

This is my second post. The first was about Ihsan in professional work, giving your best when no one is watching. This one is about a different challenge: not being deceived by what sounds right. Both come back to the same thing, sincerity. May Allah protect us from being impressed by our own cleverness, or anyone else’s.

References

Al-Ghazali, Ihya Ulum al-Din, on the etiquette of debate and the evils of argumentation. Al-Ghazali, al-Munqidh min al-Dalal (Deliverance from Error), on judging truth by its professors. Sahih al-Bukhari 5767, “Some eloquent speech has the effect of magic.” Quran: Al-Baqarah 2:204, speech that pleases but lacks substance.