abstract-work-in-progress

As generative AI becomes normalised within academic practice, how might these systems be integrated into creative research in ways that make thinking more visible rather than less? This practice-based PhD asks what alternative relationships with these systems might look like, and what they might reveal about how intelligence, knowledge, and learning are currently being shaped as commercial LLMs become embedded within higher education.

The research begins from the observation that these systems often compress process, producing polished outputs that can obscure the reflective, exploratory, and unfinished aspects of learning. As these tools become infrastructural to academic work, questions of who controls the means of thinking, and on whose terms, also become urgent.

Rather than critique these tendencies from the outside, I investigate them through my own doctoral practice, using an autoethnographic approach informed by my position as both learner and educator. If AI-mediated knowledge work risks making process invisible, then the methodological response is to build infrastructures that make it visible again, and to keep ownership of those infrastructures close to the researcher.

The core artefact is PhD-Live, a public digital research environment designed to keep knowledge-in-process visible. Alongside it, a suite of locally hosted AI systems, including a Supervisor Bot, Study Companion, and shared knowledge layer, creates an experimental infrastructure for exploring alternative human-AI research relationships. These systems are treated not only as tools but also as research materials and sites of inquiry.

Through this making, findings emerge from practice rather than planning. These include the challenge of building AI tools that genuinely challenge the researcher rather than reinforce existing thinking, a tradeoff between technological sovereignty and model capability, and questions about liveness as a research methodology. Together, the project explores how creative research might remain reflective, situated, and open-ended within an emerging landscape of AI-mediated knowledge production.

What can AI add to the processes of learning and knowledge-making, while preserving the boundary around the messy, generative thinking so central to creative work? And what assumptions about intelligence, agency, and knowledge are embedded in the commercial AI systems now entering educational contexts?
The research proceeds through building: developing speculative tools, live systems, and research infrastructure as primary modes of investigation, where each artefact operates as both a technical object and a theoretical proposition.

PhD-Live sits at the centre of this work: a live, public-facing digital research environment that treats the process of doctoral thinking as a core output rather than just the means to an eventual end state. Around it, a wider infrastructure is taking shape, including a ‘supervisor bot’ whose development has surfaced a central design tension: how do you build an AI tool that introduces genuine epistemic friction rather than simply reflecting the researcher back at themselves? This question points toward a principle that emerged through making rather than planning: that LLM reasoning and human associative thought operate with different logics. LLMs are built to resolve and optimise; creative research thinking often needs to stay open, contradictory, unfinished. Designing with both means holding that difference deliberately, rather than letting the efficiency logic of the tool quietly shape the thinking.

The research sits at the intersection of speculative design, live coding practice, and critical AI studies. It is shaped by its moment: the politics of knowledge, precarity in higher education, and the institutional pressures that surround the integration of AI. Within this, several threads have emerged as central: what a locally-based AI infrastructure might offer creative research practice; the distinction between liveness and performance; and how generative AI can participate in creative thinking without speaking for the researcher.

The research sits at the intersection of speculative design, live coding practice, and critical AI studies, and it treats building as its primary method: each artefact is both a technical object and a theoretical proposition. It is shaped by its moment, by the politics of knowledge, precarity in higher education, and the institutional pressures surrounding the integration of AI. Three threads structure the contribution. The first is what a locally based AI infrastructure can and cannot offer creative research practice, including where local models fall short. The second is the distinction between liveness and performance, which the work treats as an open conceptual question rather than a settled one. The third is how generative AI can participate in creative thinking without speaking for the researcher. Together these form an account of building and learning with and through AI systems while holding open the unfinished, exploratory thinking that creative work depends on.

The research sits at the intersection of speculative design, live coding practice, and critical AI studies, and it treats building as its primary method. It is shaped by its moment, by the politics of knowledge, precarity in higher education, and the institutional pressures surrounding the integration of AI. Three threads structure the contribution. The first is what a locally based AI infrastructure can and cannot offer creative research practice, including where local models fall short. The second is the distinction between liveness and performance, which the work treats as an open conceptual question rather than a settled one. The third is how generative AI can participate in creative thinking without speaking for the researcher. The originality of the work lies not in occupying a gap in the literature but in treating building with and through AI as a method for investigating the conditions under which knowledge-making happens within a creative institution.

—

what happens as generative AI use becomes more normalised within our academic settings? This research investigates from this starting place, questioning what this means in regard to our understandings of intelligence, knowledge and sovereignty in the face of growing integration of these commercial AI systems into many avenues of our thinking and production. This tension is something that the research aims to investigate.

Assuming these circumstances as a foundational backdrop, this research concerns itself with investigating core questions situated in this reality. Framing my research as a central site within this, the infrastructure of this research itseld becomes a place to examine these questions.

What happens as generative AI usage becomes normalised within our academic settings? This practice-based research investigates what this normalisation means for our understandings of intelligence, knowledge, and sovereignty as commercial AI systems are integrated into creative thinking, learning, and production.

These AI systems carry embedded assumptions about what intelligence is, what learning looks like, and what counts as knowledge, and as such these assumptions are now entering educational contexts rapidly %% is this a bit repetitive? %%. The kinds of knowing creative practice depends on, tacit, exploratory, unfinished, are exactly the kinds these systems struggle to support%% i dont know if struggle to support is right, as much as they werent necessarily designed with this in mind %%, and the institutional adoption of commercial AI is reshaping creative academic work before the implications are fully understood

The research investigates these questions from within. My own doctoral practice is both the site and the subject of the investigation, and the research infrastructure built to support that practice is the place where the questions are examined materially. This infrastructure includes PhD-Live, a public digital research environment that publishes the working process of doctoral thinking as it happens, and a local AI infrastructure of bots and a shared knowledge layer that function as alternatives to commercial systems. Each component operates as both a technical object and a theoretical proposition: a way of building with and through AI rather than only using or critiquing it.

Through this making, the research surfaces findings that emerge from practice rather than from planning: an echo chamber problem in AI tools that hold deep researcher context, a capability gap between local and frontier models that names a sovereignty/quality tradeoff, and an open question about what a commitment to liveness, keeping knowledge public whilst still forming, might offer as a model for learning in an era of AI-generated outputs.

—