There is a stern warning to anyone who wants to study Kabbalah, a mystical tradition within Judaism that seeks to uncover the hidden nature of God and the universe: Just wait. In traditional practices, prospective students are urged to delay their studies until they reach the age of 40, at which point it is assumed they will have developed the inner ballast to withstand what they might encounter.

We should consider doing something similar for artificial intelligence.

I understand how extreme this may sound, especially now that artificial intelligence has already entered seemingly every area of our lives. But the extremity is the point.

We have been so quick to treat access to new technology as an inevitability that we have failed to question whether every powerful tool should be placed in the hands of every developing mind the moment it becomes available.

I also see this as more than a theoretical issue. In the graduate program where I teach brand strategy, my students recently completed thesis projects that involve repositioning brands or concepts that have fallen out of pace with culture. At the beginning of the school year, we asked them to sign a document acknowledging that they understood the rules governing their use of artificial intelligence: They were allowed to use A.I. for research and for the visualization of ideas, but they were instructed not to use it to replace original thinking and writing.

We reminded them of these rules again at the beginning of the thesis process, because the thesis is not simply a deliverable or an exercise in presentation, but the place where their thinking is supposed to become visible. Because brand strategy is so dependent on the compelling communication of ideas, this is the part of the program in which students have to wrestle with language, evidence, doubt and the responsibility of making an argument they can confidently stand behind and defend.

Despite the agreement they signed and several serious reminders, we suspected that some students had indeed used A.I. in place of their own original writing. When we asked, they squirmed and then admitted they had used A.I. to augment and synthesize their writing. I found their use of the word “augment” especially revealing because it allowed the act to appear smaller than it was; it made the intervention of A.I. in their work sound cosmetic and clerical, when what was actually being altered was the student’s relationship to struggle, authorship and accountability.

I don’t think my students are inherently deceitful or that they set out to behave with contempt for the work, the program or their faculty. While I was grateful that they readily admitted to the infraction, I wonder if they understood the gravity of the situation. This is what scares me most, as it is something far more ordinary: Students seem unable to resist the ease that L.L.M.s provide them to wrestle with and craft the construction of their ideas, and are incapable of tolerating the discomfort of searching for language when language can be produced for them. They also seem unwilling to stay inside difficulty long enough for the difficulty to teach them something they will likely never learn any other way.

There is something druglike about using A.I., not because it necessarily produces pleasure, but because it removes a particular kind of pain. A person who grows up with A.I. may never know the necessary agony of a blank page, the false start, the dull vocabulary and the private embarrassment of a first draft. A student who relies on A.I. too early will likely never sit long enough with confusion to distinguish it from failure.

Confusion is a necessary state wherein they learn how to think for themselves. (Interesting turn of phrase, isn’t it? As though a person could think for anyone else.) Language is not merely how we express ourselves; it is one of the ways we become ourselves. We write in order to find out what we think, we revise in order to discover whether or not we believe it, and we speak, stumble, clarify, retract and try again and again in order to understand the space between ideas and truth.

The Jewish tradition treats Kabbalah as knowledge that requires readiness. Actual wisdom doesn’t come instantly, and it certainly doesn’t arrive fully fluent. Wisdom is the slow education of the self, built by lived experience. Often we resist it, and more often, it requires us to revise what we once thought we knew. This can be humiliating, but that is why wisdom is also associated with growth. There is a reason we were told “No pain, no gain” and “I learned this the hard way” over and over as children.

I will admit, for a mature person, A.I. can be useful. A formed mind can argue or wrangle with the machine, reject its false polish and recognize when the language it constructs to redescribe the world has become too smooth or soulless. Someone who has lived with words long enough can sense when a sentence has not been earned, and someone who has already developed a voice can use the machine without allowing the machine to become the voice.

An unformed person may not know the difference, and that is also where danger lies. The obvious problem is cheating. But cheating is only the most visible and least interesting part of the problem. The more profound risk is that young people will use artificial intelligence to bypass the development of their own interior lives and fail to hone their actual intelligence.

We do not wait for children to turn 16 to hand them car keys because cars are evil, and we do not insist that surgeons train for years because scalpels are immoral. We build thresholds around powerful things because power requires judgment, and judgment can’t be downloaded or installed. Before a person asks a machine to write for them, they should know what it feels like to write from the depths of their own uncertainty and identity. Before a person lets a machine imitate their voice, they should have endured the long, uneven, often humiliating process of acquiring one.

Of course, a rule like this is impossible as policy, and surely 40 is very late to adopt A.I. I know that artificial intelligence will not be held back by a cultural admonition, a syllabus clause, a signed agreement or even a teacher’s plea. There will be no collective vow of restraint, because the technology is already woven into the systems in which we are learning and working.

But impossibility does not make the argument meaningless. Sometimes a rule is valuable not because it can be enforced, but because it tells us what a culture reveres. The old caution around Kabbalah suggests that some knowledge is not improved by premature access. Kabbalah reveals that depth requires preparation, that mystery should not be consumed as content and that the human approaching any text matters as much as the text itself.

We need a similar caution now, not because artificial intelligence is mystical or sacred, but because it enters us through language, and language is the medium through which we know ourselves.

Debbie Millman is a designer and brand strategist, the host of the podcast “Design Matters” and the chair of the Master’s in Branding program at The School of Visual Arts.

The Times is committed to publishing a diversity of letters to the editor. We’d like to hear what you think about this or any of our articles. Here are some tips. And here’s our email: letters@nytimes.com.

Follow the New York Times Opinion section on Facebook, Instagram, TikTok, Bluesky, WhatsApp and Threads.