A Posthumanist Approach to AI Literacy

Synthesis: Wang and Wang (2025) argue for a posthumanist reframing of AI literacy, moving beyond the humanistic view of AI as a discrete "tool" used by autonomous human agents toward understanding AI literacy as an understanding of how meaning emerges through the entanglement of human and AI agencies. Through a case study of two multilingual undergraduate students (Zhimo and Asuka) in US writing courses, they document a productive tension between students' experiments with posthumanist literacy and their entrenched humanistic assumptions, showing how posthumanism offers a relational approach to cultivating AI literacy in language and literacy education.

Key Findings

• The study contrasts two distinct approaches: Zhimo embodies a critical-pragmatic humanistic AI literacy, conceptualizing ChatGPT as a sophisticated "tool" (调教 / "tiao jiao", to train and manipulate) that he directs, while Asuka adopts an anthropomorphic posthumanist stance, treating AI as an agentic collaborator whose authorial agency is entangled with her own.
• A central tension: Zhimo consistently asserts his human authorial agency while his writing process is nonetheless deeply entangled with AI — a posthumanist reading reveals that his "fix my language but not my meaning" approach still co-produces meaning through human-AI intra-actions.
• Posthumanism de-centers humans as the sole autonomous meaning-making agents; agency is "enacted" through intra-actions between bodies, ideas, materials, language, technologies, and space (drawing on Barad, Latour, Bennett, Deleuze & Guattari's assemblage and rhizomaticity).
• The framework repudiates both extremes: uncritical anthropomorphization of AI (which encourages overreliance and a transactional view of literacy) and dismissing AI as a mere tool (which leads to punitive measures that undermine students' exploratory meaning-making).
• The article proposes that a posthumanist approach to AI literacy is essentially relational rather than transactional work, questioning celebratory corporate anthropomorphism and the anthropocentric discourse of human subjects leveraging AI "tools."

Study Design & Method

The study is a qualitative case study of two multilingual undergraduate students — Zhimo (a first-year student from China in an academic writing course) and Asuka (a Japanese student in an elective writing-intensive course on generative AI and writing) — at a private US college. Data were collected in Spring 2024 through 1.5–2 hour semi-structured interviews, students' AI-assisted writing artifacts, and guided reflections. A thematic analysis examined the students' AI-mediated literacy practices through both Wang and Wang's (2025) critical AI literacy model (awareness, positionality, human-AI interactions, evaluation of AI affordances) and posthumanist theory, triangulating data sources to understand how the students conceptualized, positioned themselves with, intra-acted with, and evaluated AI.

What this means for practice

• Instructors. Have students interrogate AI-generated text as a co-constructed, contingent artifact rather than a static output: ask whose agency produced each revision and what the system's word choices foreclose for the writer's rhetorical purpose.
• Instructors. Extend AI literacy beyond textual production to the algorithmic and linguistic dimensions of meaning-making, so that multilingual writers can trace how a system's standardized English reshapes the claims they are trying to make.
• Instructors. Build legitimate refusal into assignments alongside AI use, treating a student's ethical decision to decline AI for a task as a position to discuss rather than a compliance failure.
• Learners. Examine whether you treat AI as a "tool" you fully direct or as a collaborator with independent intent: Zhimo's case shows meaning was still co-produced even while he believed he was only fixing language, as agency was distributed across the exchange.

Limitations

The study is a small qualitative case study of two students in one institutional context, bounding generalizability. The posthumanist theoretical framing is interpretive and does not offer measurable learning outcomes. The authors acknowledge a posthumanist approach is not an all-encompassing framework and must be situated within broader unresolved debates about AI's societal and ethical implications. The case-study design emphasizes depth of understanding over breadth.

Citation

Wang, Z., & Wang, C. (2025). A posthumanist approach to AI literacy.