By Carly Schnitzler, Annette Vee, and Tim Laquintano
A little over four months after the publication of TextGenEd: Teaching with Text Generation Technologies, we’re thrilled to highlight another 16 open-access assignments that enable teachers to integrate text generation technologies into their courses and respond to this crucial moment. TextGenEd was perhaps the first pedagogical collection explicitly focused on teaching writing with text generation technologies, but Continuing Experiments joins a rich ecosystem of open access collections including Harvard MetaLab’s AI Pedagogy Project and the MLA-CCCC Joint Task Force’s Exploring AI Pedagogy. Here, we publish the first of our planned semi-annual addendums: Continuing Experiments is a sustained response to the constant change that is built into teaching writing with text generation technologies. Amid the breakneck speed of AI developments, teachers are finding their footing in their classrooms by practicing, experimenting, and even resisting the tech in innovative ways.
One striking difference in the submissions for Continuing Experiments was the number of assignments incorporating prompt engineering. The call for the TextGenEd came out prior to the release of ChatGPT and the assignments reflect a range of work with text generation technologies from Markov chains to RNNs to LLMs. But almost all of the submissions for Continuing Experiments focused on LLMs, in particular ChatGPT or other openly accessible chat-based LLM interfaces. And we saw a new category emerge in prompt engineering: assignments that support students learning to iterate on their interactions with chatbots, honing the output towards particular goals. From logs that record their learning processes (Gogan) to reprogramming outputs to better reflect world Englishes (Gupta), this new category of assignments help students collaborate with the models in their current dominant interface.
Other assignments continue the work of TextGenEd in developing students’ AI literacy, with a renewed focus on the ethical and cultural impacts of AI as a critical in understanding these tools (Gallagher). In the creative explorations category, resources like AI Dungeon (Eldin) and Midjourney (Hutchinson and Jensen) expand the repertoire of crafting and analyzing creative narratives with text and image generation tools. The reflections the assignment in our ethical considerations category makes are philosophical (Proulx), examining the question “How can we use Generative AI as a writing tool and still speak for ourselves?” through a specific disciplinary lens. Professional writing assignments demonstrate and probe the utility of LLMs in applications from transcription (Horton) to resume creation (Large). Students consider how computational machines have already and will become enmeshed in communicative acts (Gillo) and how we work with them to produce symbolic meaning (Gordon) in the rhetorical explorations category.
As in TextGenEd, each assignment is licensed CC-BY-NC and includes an abstract that provides a brief introduction and outcomes plus an “assignment in brief” table that includes Learning Goals; Original Assignment Context; Materials Needed; and Time Frame. These brief summaries and the generous licensing are meant to facilitate ease of use for teachers across diverse curricular contexts. Please make these assignments your own!
While there’s good reason to resist the speed of AI development on the technological side, we believe in rapid publication of pedagogical responses. These assignments have undergone an accelerated editorial review and drafting process to showcase some of the thoughtful, of-the-moment experimentation happening in writing-intensive classrooms across disciplines.
As the name implies, we will continue to publish and share Continuing Experiments as text generation technologies and teaching evolve and inform one another. As you experiment and adapt these assignments and others in your own classrooms, please consider submitting to future editions! More information on our submission and publication processes can be found here. For now, we are so excited to share this latest collection of assignments with you.