"Instead of a prescriptive approach to AI-assisted writing and plagiarism, assignments like this try something more proactive, allowing students to experience and then articulate for themselves the issues at stake. Furthermore, this approach moves students beyond antagonistic discussions of plagiarism to consider the potential uses of AI as a writer’s tool. As Grant Otsuki argues, there’s no putting the AI genie back in the bottle and little sense banning its use outright. Instead of teachers ‘pretending AI doesn't exist’ or treating it merely as an antagonist, Otsuki encourages us to bring it into the classroom: ‘it might be time to train people to write with AI’ (2020). What would that look like? While many writers are familiar with various assistive technologies, there are practically no examples for training writers with text-generating AI and language models like GPT-2. Thus, the research questions from my experiment also explore students’ strategies for AI-assisted composition. How might we write with these tools? What skills might writing with AI newly require? How does this challenge our assumptions about textual communication? And what potential risks and harms must we navigate? ‘Cheating’ offers a starting point, but it opens onto much more complex and consequential topics, from human and nonhuman agency to the threats of disinformation and algorithmic bias. Thus, the assignment moves students from a familiar ethical context (is this plagiarism?) to new ethical questions (whose writing is this? what effects does it have?) and to new forms of writing practice (can I collaborate with AI?). The resulting insights can join a much broader conversation about AI literacy, including whether or how we might ethically and productively work with AI.”
No comments:
Post a Comment