Editorial: Generative artificial intelligence and writing instruction in K–12 and college education

Abstract

In the call for papers for this special issue, we noted that generative artificial intelligence has profoundly changed how teachers and students approach writing tasks across the globe. Some researchers have documented benefits of using GenAI in writing instruction (e.g., Bueie et al., 2026;Graham et al., 2026;Nguyen et al., 2024;Tate & Warschauer, 2026). However, other researchers (e.g., Albadarin et al., 2024;Baron, 2023;Gow, 2024;Lee et al., 2024;Loos & Radicke, 2024) have raised concerns about how the use of GenAI for writing might diminish writing skills, encourage cheating, and introduce bias and false information into texts. There are also growing concerns about how extensive use of GenAI may negatively impact the environment (Zewe, 2025).The many user interfaces and rapid development of GenAI make it a moving research target. OpenAI's ChatGPT, for example, has undergone multiple transformations since its launch in 2022. These transformations include the introduction of visual user interfaces as well as backend aspects such as increasingly sophisticated language models. Researching how GenAI is and can be (mis)used in educational contexts will continue to be important as long as this developmental pace continues. Teachers and policy makers must be able to make informed decisions regarding how or how not to integrate new forms of GenAI in schooling. In Norway, for example, reports from researchers and teachers about how students use GenAI has led the Norwegian government to enforce safe exam browsers in national exams.Research on GenAI is also urgently needed because most chatbot providers are private companies that resist disclosing detailed information on user activity. It is perfectly plausible to think that companies who provide both GenAI products and, for example, word processing software are able to provide detailed analyses of how writers prompt AI and use the output in digital writing. As long as such information remains undisclosed it is imperative that researchers continue to explore and describe how GenAI shapes writing processes and writing instruction, including the assessment of writing.GenAI and writing will continue to be an important research topic because the use of GenAI for writing may, in the long run, change how we understand what it means to be become a writer, and, in turn, how schools can effectivelly prepare student writers to thrive in an era of (potentially) transformative technological change.Together, the papers in this collection offer in-depth investigations into spontaneous teacher and student use of GenAI as well as insights into effects of design-based use. The collection is comprised of three broad categories: teacher use and perception; student use and perception; and feedback and revision.Four papers relate to teacher use and perception of GenAI in writing instruction. Goldman et al. (2026; https://doi.org/10.3389/feduc.2025.1710974) surveyed 420 US high-incidence special education teachers and found them to rarely utilize GenAI at all in writing instruction. They did note, though, that use was associated with teachers' attitudes toward GenAI and recommend professional development to increase teachers' ability and willingness to utilize GenAI in instruction. A Chinese survey of 1,220 teachers in K-12 (Ma & Zhang, 2026; https://doi.org/10.3389/feduc.2026.1755301) found teacher use to be high (98.3 % of participants reported to use GenAI) and that perceived usefulness of GenAI was one important predictor. Another predictor was age with older teachers less inclined to use GenAI. These researchers, too, suggest professional development as a means to boost the use of GenAI. A third survey (Barghi et al., 2026; https://doi.org/10.3389/feduc.2026.1779680), from Türkiye, included responses from 221 Turkish teachers of English as a foreign language. The researchers found that 71.5% of the respondents used GenAI to, for example, plan lessons and assess texts. Those who did not use GenAI tended to cite lack of training as one of the most important reasons and the researchers recommend structured professional development to reduce the proportion of teachers not using GenAI. The last paper related to teacher use and perceptions (Jeffery et al., 2026; https://doi.org/10.3389/feduc.2026.1792351) presents an investigation of 129 Norwegian teachers' perceptions text features associated with students' generative AI use. Teachers reported several features, such as lack of voice, incorrect content or shift of style from previous student texts. The researchers note that the findings imply that teacher's specialized knowledge of their curriculum, their subject, and their students will remain crucial to their writing assessments.Two papers in the collection related to student use and perception. Brynhildsen (2026; https://doi.org/10.3389/feduc.2026.1767132) used video data to explore how ten students interacted with chatbots during writing. The results showed that the students differed in their use of GenAI with some students outsourcing writing to GenAI and some being more reluctant to use it. Bakken et al. (2026; https://doi.org/10.3389/feduc.2026.1800040) investigated how GenAI could be used by primary school students to facilitate transnational collaboration. The researchers found that the process enabled students to engage with and negotiate the roles of author, reader, and editor, while simultaneously attending to questions of textual quality, meaning, and audience in the translated texts. Three papers, finally, related to feedback and revision. Li and colleagues (2026; https://doi.org/10.3389/feduc.2026.1736489) examined how well ChatGPT 4.1 could evaluate revision patterns in young students' argument writing and found the results to be compelling. Westby et al. (2026;https://doi.org/10.3389/feduc.2026.1740044) investigated the effects of an adapted chatbot that would only ask questions, not provide students with writing, when students engaged in prewriting activities. They found the constrained chatbot to be able to support students in the prewriting phase of classroom writing, even though some students needed extra attention from teachers and peers. Helm and Hesse (2026; https://doi.org/10.3389/feduc.2026.1737037) presents a model of how GenAI can be integrated in feedback to students. They argue that GenAI may well serve as a low-threshold resource in the early stages of writing. However, echoing Westby to some extent, they also caution that AI cannot replace human readers and that making effective use of GenAI-generated feedback requires strong reading skills.The papers in this collection demonstrate how GenAI might be utilized productively, but also highlight obstacles like poor preparation to use AI tools effectively, the voiceless quality of AI-generated text, and the realistic fear that GenAI may prompt students to completely outsource writing. Taken together, the collection suggests that GenAI will remain a contested tool for writing instruction and writer development in schools, and that writing researchers are only beginning to understand the potential opportunities, affordances, and challenges it presents.