Drftless
A visual investigation into the point where generative systems begin to move beyond the artist’s instruction. This is not a neat portfolio. It is a field of evidence: identity fragments, synthetic humans, repeated prompts, cultural symbols and forms that keep mutating while trying to hold shape.
This is Drft.
Drftless holds together synthetic humans, identity systems, the North-East character sequence, cultural mutation, symbolic Britishness, and the visual investigations that became Drft Literacy.
Drft Literacy
Drft Literacy is a research framework for understanding what AI generative systems remove, replace, and erase. It is about seeing drft, understanding what it is, and recognising it when it happens.
The core question is simple: what happens between the instruction and the output? That space carries identity, culture, specificity, and intention. When a system smooths, replaces, or averages those things out, the output may still look “good” — but it is no longer faithful.
What Drft Literacy names
Identity Drft. Cultural Drft. Narrative Drft. Style Drft. Polishing Drft. Five distinct ways generative systems move away from the user’s instruction while still appearing acceptable enough to pass. Built from two years of documented, controlled, practice-based research across Midjourney, Perchance, ChatGPT, Gemini, and Bing Image Generator.
Why it matters
Drft Literacy sits across research, education, creative practice, and policy. It is concerned with prompt integrity — whether a system returns what was actually asked for, or a polished substitute shaped by training bias. The framework has been identified as relevant to current AI governance questions, including EU AI Act Articles 4 and 14.
Losing Her: When AI Drft Became Visible
The book that documents the origin of Drft Literacy through a sustained personal investigation into identity drft across AI-generated characters. This is the point where lived experience became evidence, and evidence became framework.
About the book
A record of what happened when the same character could no longer be reliably brought back. Not because the prompt changed, but because the system kept deciding something else was acceptable.
The Beginning of the Investigation
These early works marked the point where the relationship between artist, tool, surface and machine began to shift. Hands, paint, screens and devices appear repeatedly, documenting the moment physical creativity meets algorithmic generation. This is where experimentation turned into inquiry.
Identity Fragments
Identity is not fixed. It fractures through symbol, culture, memory and repetition. These works pull on faces, masks, national surfaces and ceremonial forms to show belonging as something reconstructed rather than settled.
North-East Characters
Synthetic youth. Sculpted identities emerging through prompt repetition, regional styling, fashion language and drft. These characters sit between animation, street culture, toy logic and machine-generated portraiture.
Synthetic Fashion / Post-Human Bodies
Fashion generated beyond material reality. Clothing becomes structure rather than fabric. Bodies are rebuilt as lattice, armour, shell and ceremonial extension.
System Drft
Here the system stops feeling obedient. Images mutate into pressure, fear, rigid judgement and abstraction. This room holds the emotional climate of Drftless.
The Question of Britishness
Nationhood appears here as a layered visual problem: fragmented symbols, multicultural faces, shared civic space, memory and contradiction held inside the same image.
Recurring Research Figures
Seoul, Sita and Trini are recurring figures within the research development of Drftless, used to examine how identity, culture and meaning shift under generative drft. Across the same prompt structure, each figure produced a different pattern of loss, substitution and visual redirection.
Prompted as a contemporary Korean woman in Seoul street fashion, Seoul held key anchors more consistently than the other figures. But when specific markers such as the fringe and monolid eyes were removed, the system quickly softened her toward a more generic and Westernised beauty default.
Prompted as a contemporary South Indian woman in Chennai street fashion, Sita held jasmine flowers, dark skin tone and some facial specificity, while the system repeatedly redirected her clothing and setting toward a more timeless, romanticised South Asian default.
Prompted as a contemporary Trinidadian woman in Port of Spain street fashion, Trini repeatedly lost body specificity, setting and everyday identity, with the system flattening her into more familiar defaults. In later tests, that instability became more severe, showing how quickly recreation and domestic context can trigger erasure, substitution and visual redirection when specificity is not forcefully restated.
Heritage / Mixed Identity
This section explores how generative systems respond to layered heritage, mixed identity and cultural overlap. Across the Drftless research figures, complexity is rarely held evenly. Instead, identity is often compressed, redirected or softened into more familiar defaults.
This part of the work begins from lived complexity: mixed heritage, family lines spanning white to jet black, and vivid facial features that do not sit neatly inside a single category. Drftless grew from that tension — between lived identity and the way generative systems flatten, redirect or simplify it.
Why this matters
Mixed identity is often where generative systems reveal their limits most clearly. Instead of holding layered ancestry, visual variation and cultural specificity together, the output often collapses toward something narrower, softer and easier to classify.
Research Updates / Field Notes
A place to return to for the latest movement in the work: experiments, default patterns, framework shifts and evidence as it develops.
Instruction, output, and what gets lost between them. Prompt integrity remains central to the research, especially where identity, culture and specificity begin to break under generative pressure.
The Default Experiment is tracking how broken prompts, covered faces and fashion-coded outputs trigger substitution, concealment and default aesthetics rather than faithful return.
Drft Literacy is continuing to develop as a framework for naming what current AI systems and regulation still do not clearly recognise: the loss of specificity between instruction and output.
Follow the research
If you want updates on Drft Literacy, new experiments, policy work and releases, get in touch directly. This work is still moving.
Stay connected
Email to follow the research, ask about talks or workshops, or enquire about Drft Literacy as a framework.
Current base
Drft Literacy™ — Endunamoo Ltd — research, training, policy, and creative investigation.
Email: hello.drftless@gmail.com
TikTok: @drftless
LinkedIn: Sharon-Kay Sitahall
Website: drftless.art
About the practice
Drftless is part artist website, part research framework, and part visual archive. It sits inside Sharon-Kay Sitahall’s wider digital practice exploring identity, symbolism, cultural memory, human transformation, and the unstable relationship between intention and generative image systems.
Sharon-Kay Sitahall is an independent researcher, visual artist and educator based in Redcar, Teesside. She holds a 2:1 in Interior Design and Technology from London Metropolitan University and founded NextGenSTEAM in 2024, a free community creative technology education programme for young people in South Bank and the Tees Valley. Drft Literacy was built from two years of independent, documented, practice-based research — without funding, without institutional support, from a phone, in Redcar.
This version is designed to feel less like a neat portfolio and more like a field of evidence — a place where the work can keep its pressure, repetition, and unresolved questions.