The arrival of AI is for the humanities not merely a technological revolution but a deeper challenge to how knowledge is produced and how subjectivity is formed. When reading, writing, interpretation, organizing materials—even certain kinds of critical thinking—can be rapidly simulated by algorithms, the value of the humanities can no longer rest only on “knowledge output.” The humanities must return to more basic questions: Where does knowledge come from? How is judgment formed? How does a person, amid language, experience, institutions, media, and bodily feeling, form the way they understand the world? In other words, what cannot be replaced in the humanities in the age of AI is not that humans write or analyze better than AI, but that the humanities can ask how writing, analysis, judgment, desire, and self are generated. Meditative humanities must push one step further. It asks not only how knowledge is produced by society, history, and institutions, but further: Before knowledge is produced, how is the person who can see, pause, question, and endure uncertainty trained? That is the core meaning of contemplative practice and koan practice in the age of AI. Meditation is not a relaxation technique; a koan is not a religious riddle. Together they form a basic training for how knowledge is generated: enabling a person, before an answer appears, to see how they crave an answer; before judgment forms, to see how judgment is pushed by language, habit, emotion, and desire; before the self rushes to speak, to see how that “I” is provisionally assembled. 1. AI can generate answers, but it cannot generate sustained inquiry (參究) The power of AI lies in producing, simulating, predicting, and optimizing. In seconds it can generate structurally complete, tonally mature text that seems theoretically deep. This forces the humanities to face a more basic question: If language, analysis, interpretation, and critique can all be simulated, what remains for the humanities? The answer cannot stop at defensive apology. The value of the humanities is not to outpace AI in volume or speed of output, but to work at a level of problem that differs from the level at which AI operates. AI can generate answers, but the humanities ask: How are answers generated? What is the ground of that generation? What kind of living state can truly hold a question, instead of immediately converting the question into an exportable answer? Here, contemplative practice and koans offer the most crucial insight. Knowledge does not begin with answers but with the capacity not to flee immediately toward answers. Genuine understanding is not the stacking of information but whether a person can stay in confusion, observe in uncertainty, and when familiar concepts fail, still refrain from hastily re‑closing the world with familiar language. Koan inquiry is exactly an extreme training in this kind of knowledge generation. A koan does not hand down a ready answer. It stops language, disables concepts, and lets automatic reaction hit a boundary. It compels us to see: what we call understanding is often a quick application of an old frame; what we call judgment is often habitual reaction; what we call “I know” is often a cover for unease. Therefore AI can generate answers, but it cannot do our inquiry for us. It can offer explanations, but it cannot bear our doubting engagement (疑情) for us. It can mimic the tone of insight, but it cannot undergo the embodied state, before insight arises, of being pressed to the edge by a question. The real depth of the humanities begins here. 2. The ground of knowledge generation: not information, but awareness The greatest danger in the age of AI is not only that students stop writing, but that they stop undergoing the process by which knowledge is generated. When answers appear too quickly, thinking may be short‑circuited. When language is too fluent, confusion may be flattened. When analysis is too complete, the weight of the question itself may dissolve. But knowledge is truly generated not after the answer is finished, but before the answer has yet formed. It happens when a person begins to feel: Why do I think this way? Why am I drawn to this answer? Why do I rush to a conclusion? Why do I fear having no position? Why must I immediately prove I am right? These questions are not psychological ornaments to knowledge; they are the ground of knowledge generation. The value of meditative training is that it lets people see this ground. When we sit and observe breath, body, thoughts, and feelings, we discover: what we call “I am thinking” is in fact countless sensations, memories, language, fear, desire, and habit running automatically. What we call “my judgment” is often not the fruit of pure reason but something pushed out amid bodily tension, emotional reaction, social expectation, and linguistic habit. This seeing is not anti‑intellectual. On the contrary, it is a deeper condition for knowledge. Without awareness, knowledge easily becomes automatic reaction; without pause, critique easily becomes mere posture; without inquiry, thought easily becomes automatic arrangement of concepts. Thus what meditative humanities propose is not to add a little inner exercise beside humanistic education, but to point anew: the ground of knowledge generation is not only information processing but the capacity for awareness, endurance, pause, inquiry, and transformation. 3. AI is not only a tool but a new environment for subject formation In the age of AI, this kind of disclosure must go further. The humanities have disclosed how language, history, class, gender, colonialism, institutions, and power shape the subject. Today the subject is shaped not only by these structures but continuously recomposed by algorithms. Recommendation systems tune our attention; language models reshape our expressive habits; data structures reinforce our preferences and judgments. When thinking becomes too fluent and answers too easy, we may unknowingly outsource subjectivity to probabilistic models. AI is therefore no longer merely an objective tool but a new environment in which subjects are generated. Truly leading humanistic questions are no longer only “how to raise efficiency,” but: As AI participates in the constitution of my thinking, what kind of being am I becoming? In the pursuit of correct answers, am I losing the ability to dwell with confusion, uncertainty, and blank space? Am I, amid fluent language templates, gradually losing the awareness of how my own intentions are generated? These questions constitute the humanistic education truly needed in the age of AI. 4. Training in meditative humanities: GBSD/KPSD subject education The meditative humanities promoted by the Vimalakīrti Institute are not a little relaxation added beside the humanities, nor a reduction of meditation to stress relief, but a method of subject education—more precisely, a GBSD/KPSD (key‑phrase structural deconditioning) training of the subject. Traditional humanities excel at analyzing and criticizing texts, history, and institutions, yet such analysis often still stops at already formed results or takes a third‑person observational stance. Meditative humanities bring critical sight back to the site where experience happens, including the “inner site”: in every moment of reading, writing, using AI, or everyday life, observing how a thought arises, how an emotion occupies, how a judgment forms, how an “I” is provisionally assembled in the craving for an answer. Therefore meditative humanities train not only the ability to understand texts but the ability to notice, deconstruct, and reconstruct the subject. So‑called “key‑phrase structural deconditioning” (GBSD/KPSD) is not simple negation of existing structures; through contemplative practice and koan it lets one see how thoughts, feelings, judgments, and self are conditioned by language, habit, institution, and situation—and, before a question is immediately converted into an answer, to remain inside the question so that the question in turn illuminates our habits, fears, and cognitive limits. The function of the koan is to interrupt the habit of automatic understanding so that we do not only handle new problems with old frames but, where frames fail, see anew how we are shaped by frames. Finally, through the stability cultivated in meditation, we no longer fully identify with our own reactions and can preserve a small yet crucial gap in the chain of automatic reaction. That is the deep meaning of meditative training in the age of AI: it opens a “breach” between reaction and action, letting tightly sealed automatic reaction begin to loosen, and turning it into a liminal space that can be felt, borne, and chosen anew. 5. The gap: the beginning of freedom and responsibility The core logic of AI is prediction and optimization. Through data it recognizes patterns and seeks the most efficient solution within existing frames. It excels at reducing uncertainty and making human behavior more predictable, manageable, and operable. Yet human freedom appears precisely in the “gap” not yet fully predicted. This gap is not emptiness and not escape. It is a space in which the world can be responded to anew. In this gap we can pause to ask: Why does this answer attract me? Where does this judgment come from? Is this desire mine, or was it summoned? Where is this choice taking me? More importantly, this gap is not only an epistemological issue but an ethical one. AI can simulate and optimize decisions, but it does not bear consequences. Humans must. The disclosure of how “I” is generated in contemplative practice is not only for seeing but so that we may further ask: Since “I” is assembled in this way, how should I take responsibility for this assembled judgment? This is what cannot be replaced in humanistic education. Genuine understanding is not only mastering information but letting knowledge enter life and transform perception, judgment, and action. AI can describe the world but cannot bear the world. It can analyze suffering but does not experience suffering. It can generate ethical language but does not bear ethical consequences. The path from grasping information to transforming cognition and subjectivity is the irreducible depth of the humanities and of meditative humanities. 6. Digital koan: beginning again at the edge of thought In the age of AI, every stretch of seemingly perfect AI text may become a “digital koan.” A koan does not hurry to supply an answer; it first interrupts our habit of seeking answers. When AI outputs fluent, precise, structurally complete text, we must not only ask, “Is this text correct?” but stop to ask: What has this fluent substitute saved me—and what has it cost me? Am I participating in the creation of meaning, or consuming a mediocre copy? Am I using AI, or being trained by AI? This is the radical meaning of the koan today: it brings us to the edge of thought and compels us to see how we depend on existing frames, language, and patterns of reaction. In this sense the digital koan is not anti‑technology but trains awareness within technology; not a refusal of generation but not losing thought inside generation; not a denial of efficiency but a questioning of the direction of efficiency. Therefore the core of AI education should not be only how to prompt, how to generate content, or how to raise efficiency, but how to maintain awareness in operation, how to recognize hollowness within fluency, and how to preserve a gap for subjectivity within automation. 7. From anxiety in the liberal arts toward “architects of meaning” and bearers of consequence For students in the liberal arts, the anxiety brought by AI is real. Yet that anxiety may also become a turning point. The future role of liberal‑arts students is not to add humanistic decoration to technology but, in an era of rapid technological expansion, to become “architects of meaning” and bearers of consequence. Through the CARE framework (concentration‑awareness‑responsiveness‑engagement)—focus, awareness, responsiveness, participation—we learn to redraw the boundaries of problems. What matters is not only what technology can do but what futures are worth creating, and what forms of human growth must not be reduced to algorithmic training and optimization. Humanists must guard spaces that are “not optimizable”: boredom, pain, failure, waiting, aimless exploration, and contemplation before an answer has formed. These seemingly inefficient experiences are where people can transform, bear, and begin again. In this sense interdisciplinarity must also be redefined. The issue is not only whether the liberal arts should learn technology but who decides the uses of technology. Interdisciplinarity without deep awareness easily becomes an appendage to technological expansion; those trained in meditative humanities may become those who redefine the problems of technology. Thus the future task of liberal‑arts students is not only to comment on technology but to raise questions anew at the center of technology, guard judgment, bear consequences, and remind the world: not everything that can be generated is worth generating; not everything that can be optimized should be optimized. 8. Conclusion: toward a meditative humanities of “mindfulness turned toward the world” Of course meditative humanities has its own risks: it may be instrumentalized as stress relief, or elitist as a spiritual luxury for the few; it may stop at inner awareness and fail to respond to the world. Therefore the meditative humanities spoken of by the Vimalakīrti Institute must explicitly move toward a meditative humanities “turned toward the world”: no longer stopping at private inner settling but turning toward response and responsibility for the world. As the Vimalakīrti Sūtra shows, Vimalakīrti does not teach apart from the world in a pure place but unfolds the teaching amid illness, the city, crowds, and contradiction. This reminds us that genuine awareness is not withdrawal from the world but, amid the world’s suffering, still keeping clarity, compassion, and the capacity to respond. Meditation is therefore not a way to retreat from the world but a training to enter the world more wakefully, steadily, and responsibly.