Active · June 2026
AI visual development workflow
Testing image generation tools across pre-production visual development: mood boards, location concepts, character reference. Question: does this accelerate decisions or defer them?
View CineStackLab
Where research becomes practice,
and practice becomes material.
The lab gathers short-form tests, prototypes, and evolving methods at the intersection of moving-image storytelling and emerging technology. It is not a showcase. It is a working document.
What this is
A working space, not a portfolio
The lab exists because the research and the teaching both require a place where ideas can be tested before they are stable enough to write about or teach from. Some experiments here will become research outputs. Some will become module content. Some will just be useful.
The underlying questions are consistent: What does AI actually do to cinematic authorship? Where does a tool extend craft and where does it substitute for it? What are the ethical positions a filmmaker needs to form, and how do we teach that formation?
Working methods
How it operates
Each experiment begins with a specific question, not a tool. The tool is selected (or rejected) based on whether it bears on the question. Results are documented as practice notes that can feed into research, writing, or teaching. Nothing here is claimed as finished or definitive.
The lab also serves the book Filmmaking in the Age of AI: A Practical Guide as a testing ground for claims made in the manuscript. If something doesn't hold up in practice, it doesn't go in the book.
Current work
Active experiments
Work in progress. Updated as experiments develop. Some of the functionality might not work as expected. Feedback and suggestions welcome.
Active · June 2026
Prompt architecture for AI filmmaking
Building prompt structures designed specifically for AI filmmaking contexts, where the ethical constraints are different and the relationship to reality matters in ways fiction can sidestep.
View The Blueprint StudioLive tool · July 2026
The Learning Designer
A visually structured approach to learning design for film and media educators. Sequence teaching and learning activities across six research-backed learning types — acquisition, collaboration, discussion, inquiry, practice, and production — and watch the balance of a session take shape as you build it. Designs export as JSON or a student-facing session guide.
Open the Learning DesignerOngoing · updated June 2026
Teaching AI ethics in film education
Developing curriculum material for teaching AI ethics as craft ethics, not as a separate module but as embedded in every production decision. Testing in module contexts at MetFilm.
View FrameCraftDeveloping · June 2026
Longitudinal skills passport prototype
Early prototype work on the skills portfolio system that forms the core of the EdD venture research. Building the architecture before the institution exists to test it.
View FilmFlowLive tool · June 2026
What AI cannot do in filmmaking
A running document of specific limitations encountered in practice. Less interesting as a critique, more useful as a calibration tool for anyone who wants to use these tools without losing their judgment. It now runs in full further down this page.
Use the calibration instrument belowTool
AI Creative Market Map
An interactive map of the AI creative tools landscape across development, production, post-production, inference, and distribution. Click any tool for an AI-generated description.
Tool
Assistant Editor Essentials
A practitioner resource for students starting out: a UK post-production facility directory, the skills houses are actually asking for, worked speculative letters pinned to year of study, and a draft-with-Claude letter generator built on the 4Ds.
Coming next
Next tool in development
A third prototype is in progress. Details land here once it is ready to test.
Flagship tool — live on this page
The calibration instrument
What AI Cannot Do in Filmmaking, running here in full: eighteen documented limitations across six clusters, and a 4Ds practice studio — Delegate, Describe, Discern, Diligence — for deciding what a tool should and should not touch. This is the pedagogy operating, not a screenshot of it. Open it full-screen if you prefer more room.
Lab log
Dated entries
A running record of what changed, what was tested, and what came of it. Newest first.
10 July 2026
Calibration instrument embedded; the Learning Designer goes live
The calibration instrument now runs directly on this page rather than behind an external link. The Learning Designer joins the active experiments: session design across six learning types, built for film and media educators, with a worked example on AI in film production included.
23 June 2026
AI Creative Market Map launched
An interactive map of the AI creative tools landscape across development, production, post-production, inference, and distribution. Insight generation moved server-side so visitors no longer need to supply their own API key.
June 2026
First public versions of the experiment prototypes
Six experiments published in one pass: prompt architecture, editorial review, ethics curriculum material, the skills passport prototype, and the calibration instrument. Publishing them unfinished is the point — the lab documents work in motion, including the parts that do not yet hold up.
A note on method
Why document this publicly
The lab is public because the research is public. The EdD methodology is built around transparency about process: what was tried, what failed, what was learned. Documenting experiments here is part of the practice-as-research commitment, not a marketing gesture.
It is also useful for students and practitioners navigating the same questions. Not as a set of answers (the answers are context-dependent and often provisional) but as a model of how a working practitioner-researcher approaches tools they do not yet fully understand.
Working on similar questions?
I am interested in connecting with practitioners and researchers exploring AI in creative practice. Get in touch.