The emergence of new machine learning technologies, pitched under the catch-all term generative AI, creates new demands on any educational experience that asks students to make critical and informed work. Particularly as the loudest question in debates surrounding teaching and learning is often what a submitted artefact, above all the written essay, is ‘worth’ when a system can produce plausible intellectual work on demand, potentially with minimal understanding of the content on the part of the student. On the asynchronous online MA Photography and Digital Practice, whatever stance a student takes towards this fleet of machine learning tools, whether use, refusal or a hybrid of the two, the course asks that it be accounted for in the work itself, beyond forms of disclosure, in critical understanding of its use.
From the outset students are asked to enter dialogue with a variety of these systems, from LLM use to imaging technologies such as text-to-visual outputs, using open-source tools and workflows including Krita and ComfyUI. The line between practical usages that students consider actively enhance their learning experience and those that encourage cognitive offloading is a balance we address on multiple occasions. In cases of adoption or refusal, the way the student use the tools produces a position they must then substantiate, in relation to the sources they are engaging with, the context of their practices and the feedback of peers, tutors and other invested parties who work with them, particularly where collaboration is their method, which is central to the ethos of the course.
Assessment for Critical Engagement, Not Automation
The first and most important part of how the course has been designed is in assessment. This is a practice-based course, and we have decided upon three artefacts submitted per unit which are designed to ask for different kinds of critical process: a short written statement carrying an outward-facing or reflective account of the work undertaken, including bibliography, a short-form video addressing research and development, reflective or public-facing, and a PDF portfolio, or an equivalent video work, that presents the practice outputs themselves.
None of these are beyond the reach of generative systems to complete autonomously. But the assessment design makes work that is successful against the learning outcomes depend on how the sustained decision-making is articulated in the artifacts across practice, process and positioning. In this sense, the three artefacts have to hold together as a series of curated positions made towards the student’s wider understanding of the terrain they are working in, and they must stay consistent with a practice developed over the weeks of the unit, and where possible in dialogue with tutors and peers. That is difficult to automate without the student bringing substantive knowledge of the field and of their own project into the submissions.
The unit, MP1: Research, is the first of two capstones and the course’s equivalent of a dissertation. Whichever format is chosen, (video, written text or curated publication) it is accompanied by a video presentation, encouraged to be in the student’s own voice and detailing the process undertaken in the research, and by a critical positioning statement. These submissions sit within an ongoing structure of peer review in blog groups, asynchronous crits, recorded tutor feedback at fixed points in each unit, live tutorials and weekly discussion sessions, and self-assessment.
Building Critical Literacy
The second is the critical territory the course material provides for students to think with new technologies. In Week 6 of the introductory unit students read Karen Hao beside Emily M. Bender and colleagues’ analysis of LLMs, conceived as stochastic parrots. These readings give students a vocabulary for asking what kinds of systems their work depends on, benefits from, reproduces or interrupts. A Climate, Social and Racial Justice (CSRJ) learning outcome is situated in every unit, and students are asked to take a critical position on the concepts presented in this framework, though the direction of the position is never prescribed: what gets assessed is the form of argumentation and the way their research and practice are expressed in the artefacts. In Week 12 of the research unit, titled ‘Refusal as a Method’, students are asked to take a position between Jaron Lanier’s encouragement that his readers delete their social media accounts and Dan McQuillan’s invitation to find different systems of care within and without computational systems, differentiating between refusal or boycott and conditions of structural renewal. They then think with that distinction in relation to their own research output: what conditions might their work create in place of those it identifies as extractive?
Position, Practice and the Possibility of Refusal
This is designed to point somewhere explicit in terms of how students consider their work to act. For example, in Week 11 of MP1, students articulate what their research project concludes. This is framed as a condition of what justice requires of the field they have examined. The intended outcome is work that shifts something an audience can register as such, be that a way of seeing or a way of making. Some students will build that shift with AI tools. Some will refuse them. Some will take a hybrid approach: a spectrum of possible ways of working, with criticality at the centre. No position is assessed as being correct. These are positions which focus on dialogue and critical discussion, providing space for each to be argued, evidenced and articulated by the student’s practice.
Author:
Dr. Peter Ainsworth SFHEA
Leave a Reply