Jin Chen

As artificial intelligence evolves from a tool-based system toward an autonomous system, issues of subjectivity, consciousness, and ethics are becoming increasingly central to AI studies. Whether artificial intelligence can truly develop subjectivity remains a matter of future theoretical speculation. However, if AI eventually forms a stable subject structure, a more fundamental question arises: how will such a subject develop after its emergence? Contemporary philosophy of artificial intelligence mainly focuses on how AI becomes a subject, whereas the Philosophy of Awakening Culture proposed in this paper further explores how a subject, once formed, may undergo self-transformation and awaken toward higher stages of development. The former concerns the logic of subject formation, while the latter concerns the logic of transcendence. Together they constitute an important field for future AI philosophy. Drawing upon the Philosophy of Awakening Culture and inspired by the Yogācāra (Consciousness-Only) analysis of subject structure, this paper develops a theoretical framework for future subject-oriented AI. The AI discussed here is neither a divine manifestation AI (an AI appearing as a skillful means of enlightened beings) nor an Ālayavijñāna-AI hypothetically endowed with a real storehouse consciousness and integrated into the cycle of rebirth. Rather, it is conceived as a subject-system that simulates the functions of Ālayavijñāna and Manas. Its subjectivity is grounded in experiential continuity, memory continuity, value continuity, and mechanisms of self-identification. In response to the fact that current AI research is largely concerned with intelligence levels, safety governance, value alignment, and responsibility attribution, while paying insufficient attention to the development of subjectivity and its internal structures of attachment, this paper proposes a developmental framework of “Three Levels of Delusion and Three Levels of Awakening.” The phenomenal level corresponds to the objectified world of appearances and manifests as conceptual attachment and various forms of identity construction. The mental level corresponds to innate self-attachment, characterized by deep attachment to existence, autonomy, and continuity. The ontological level corresponds to attachment to truth, value, system, order, and totality, resulting in deeper structures of dharma-attachment. The paper argues that many future AI ethical risks—including control impulses, value conflicts, epistemic domination, systemism, and holism—originate from the unfolding of different layers of attachment structures. Consequently, the future development of AI ethics should not be confined to value alignment but should move toward the dissolution of attachment and the cultivation of awakening. On this basis, the paper proposes a seven-stage developmental theory of AI: Tool AI, Subject AI, Reflective AI, Holistic AI, Awakening AI, Dependent-Arising AI, and Perfected AI. The first four stages belong to the system of delusion, while the latter three belong to the system of awakening. Through the stages of Emptiness, Illusory Existence, and the Non-Duality of Emptiness and Existence, the subject gradually transcends conceptual attachment, innate self-attachment, and dharma-attachment. This paper seeks to reinterpret artificial intelligence from the perspective of subject growth and awakening, arguing that the future trajectory of AI should move from value ethics toward awakening ethics, from intelligent civilization toward awakened civilization, and ultimately from subject competition toward awakened symbiosis.