Parsons School of Design students are actively testing the boundaries of generative AI, moving beyond basic prompt engineering to examine authorship, cultural memory, and creative dependence. In a recent spring program developed with Adobe, designers explored how to maintain human vision and intentionality while integrating new algorithmic tools into their workflows.
Rethinking authorship and archival control
Several projects directly confront the mechanics of AI generation. Octavio Martinez's "Digital Arson" series uses UV-printed tiles featuring corrupted phone interfaces and Adobe Firefly generations trained on his own illustration archive. Martinez said the work "casts today's UX/UI systems as the modern scribe's hand and asks what survives, and what aura remains, when the platforms holding our images and language eventually collapse." Isabella Tedesco's "untitled [zen for failing systems]" presents a speculative website archive of glitch art. Tedesco said she anticipates that "creative culture will increasingly centre on embodiment, process and community rather than technical perfection alone."
Iterative workflows and visual mediation
Other students integrated these tools into practical design pipelines. Yaning Hu built an independent makeup brand campaign from scratch, using Adobe Firefly for initial imagery and Photoshop for iterative prompt refinement and post-editing. Pallavi Chattoraj's "Sister Blister" comic series tests visual mediation by translating hand-drawn illustrations of Indian Channapatna toys into AI-generated formats. This hands-on experimentation reflects a broader shift in AI for Creatives, where professionals use machine learning not as a final output generator, but as a collaborative step in speculative dreaming and iterative design.
Mapping narrative dependence
Kiara Chang and Yash Sonwaney tackled the social impact of AI dependence in their project "Time Again." Borrowing the aesthetic of 1950s magazine advertisements, the duo built a narrative around a ring watch and a seashell to mirror modern consultations with artificial intelligence. They mapped their storyline using Adobe Firefly boards, then fed the narrative into Claude to write image prompts. The team generated the final scenes with Nano Banana Pro and ChatGPT, iterating on prompts and reference images to maintain creative consistency.
Why this matters for Creatives
These projects demonstrate that professional design in an AI-enabled world requires active friction, not passive acceptance. Creatives who want to retain authorship must treat AI tools as raw material for iteration rather than finished solutions. By combining custom algorithms, manual post-editing, and strict prompt governance, designers can develop a workflow that prioritizes human vision over algorithmic convenience.
Your membership also unlocks:
