Abstract: Due to the powerful algorithmic reasoning capabilities, generative artificial intelligence intervenes in human cognitive processes, leading to new forms of discipline over humans, including the alienation and dissolution of humans as cognitive subjects, transforming knowledge generation from human creative labor into algorithm-driven probabilistic games, and establishing a more concealed, microscopic, and automated "super panopticon" disciplinary mechanism. At the same time, generative AI also holds the potential to "deterritorialization" education: the "decentralization" of knowledge production can break the authority monopoly of traditional knowledge production, the reconfiguration of the teacher-student relationship can help to bring out students' subjectivity, and the emergence of "smooth" learning spaces makes educational system innovation possible. Generative AI can be both a disciplinary device that alienates humans and a highly effective deterritorializing tool. Possible "escape lines" for humanity in the face of generative AI include constructing an educational philosophy paradigm of "Becoming - Other", emphasizing embodied experience to restore the meaning world obscured by technology, viewing generative AI rationally to be vigilant against the potential risks of algorithmic discipline, and learning to use generative AI tools critically to unleash their positive potential.