2024 ASEE Annual Conference & Exposition

This theoretical perspective paper considers the affordances and risks of writing with(in)/through Generative AI (GenAI) technologies in engineering. Literacies mediate identity formation (Gee, 2015) and can serve as a weapon or shield by forcing homogenization or protecting space for alternative identities and discourses (Collins and Blot, 2006 p.122). The effects of GenAI on the development of engineering students’ professional identities have not yet been determined. However, there are some clear affordances and likely negative impacts of the widespread use of this technology for students who use named languages other than English and/or non-standard dialects. GenAI can reduce the inequitable language labor required for individuals who must translate their research and writing into other named languages or discourses, an inequity prevalent in modern STEM writing (Clavero 2011). On the other hand, GenAI can be a weapon that reinforces homogenization by generating prose in standard English by default, or what Baker-Bell (2020) calls White Mainstream English (WME), thereby reducing linguistic diversity.

GenAI represents a shift in how we write with(in)/through literacy technologies by repositioning the human in a collaborative relationship with the machine (Duin and Pederson 2021; Robinson 2022). While we have long used robots like Grammarly and spell check  to help us write (Hart-Davidson 2018), writing with GenAI has shifted from working in human-machine assemblages to revising human-generated texts to collaboratively writing text that is partially autonomously generated by machine logics. This hyper-extension of human-machine cyborg assemblages (Harraway 1991; Hayles 1999) calls into question how we understand authorship and agency (Robinson 2022). The power of GenAI as a writing tool is based on its large training data set; however, that apparent diversity belies the primacy of language practices from younger, white, more affluent users in the training data (Bender et al. 2022). Thus, when we consider that GenAI technologies primarily generate language that aligns with WME language practices (and do so better than any other language or dialect), the consequences of GenAI in such assemblages suggests a further and invisible entrenchment of WME language practices in STEM communication and a setback for efforts towards linguistic justice.

This paper first offers a theoretical definition of collaborative writing that emphasizes how agency in collaboration is shaped by interactions between human and non-human actors, including the emergent text co-written by the collaborators (Cooper 2018; Boyle 2018; Duffy 2020; Walker 2015). When we consider GenAI through this perspective, we can see how the repeated practice of writing with GenAI recursively shapes the writing practices that emerge for human agents. Through that serial repetition, GenAI will shape the future potentials for action available to individuals. Second, the paper illustrates how GenAI can shape future action and meaning potentials for engineering students by considering the nominalization of verbs in technical communication (Halliday 2004) with the grammar of animacy in Anishinabe (Kimmerer 2013). This example shows style is more than just identity politics, but provides potentials for meaning that can substantively impact the development of alternative approaches to engineering work.