Field notes from teaching AI + creativity: documenting what we're building, what we're testing, and what students teach me along the way.
JSON prompting is sold as a way to make image and video models obey. The documented evidence says something narrower and more useful: it is a way to control your own process — isolate one variable, change one field, keep a record of what you decided.
A teaching handout on the four formats you write to an AI in — plain text, Markdown, JSON, HTML — what each one is good at, and how format choice changes what a coding agent can do, verify, and report back.
Un document pédagogique sur les quatre formats dans lesquels on écrit à une IA — texte brut, Markdown, JSON, HTML — ce que chacun sait faire, et comment le choix du format change ce qu'un agent de code peut exécuter, vérifier et rapporter.
Le prompt JSON est vendu comme un moyen de faire obéir les modèles d'image et de vidéo. La documentation dit quelque chose de plus étroit et de plus utile : c'est un moyen de contrôler son propre processus — isoler une variable, ne changer qu'un champ, garder trace de ce qu'on a décidé.
A long-form field guide on the 2026 shift from generative AI as image producer to agentic AI as a controllable layer inside Blender, Figma, Resolume, TouchDesigner, Remotion, and the creative toolchain at large.
Recensement structuré des plateformes, API, workflows et sites indépendants qui exposent publiquement Seedance ou annoncent un accès Seedance. Catalogue de littératie, pas une liste de recommandations.
A structured catalogue of platforms, APIs, workflows, and independent sites that publicly expose Seedance or claim Seedance access. A literacy catalogue, not a recommendation list.
A risk-aware educational reference cataloguing free, freemium, trial-based, and open-source AI video generation tools. Not a recommendation list and not an official Concordia University resource.
Asked what students should learn to stay relevant in an AI-driven future, Nina Beguš points to the humanities — language, literature, history, philosophy, interpretation, cultural understanding.
AI's intelligence isn't a property of architecture or compute — it's an inheritance from human social complexity. And deployment is quietly eroding the conditions for its own advancement.
As generative AI raises urgent concerns around appropriation, authorship, and ethics, the question is not whether Fine Arts should ignore it, but whether artists should be excluded from shaping how it is understood, challenged, and used.
Why this blog exists: to think out loud, clarify, and maintain a proactive presence on creative AI.
Questions, reactions, enquiries — all welcome.