Labojohar Arts And AI

A prompt often starts with something too broad to picture clearly: “an artwork about Somaliland,” “a landscape from the Horn of Africa,” or “a traditional scene in a modern style.” Such phrases contain a subject but not yet a composition. They do not tell the image generator where the viewer stands, what occupies the foreground, how the light falls, or which details distinguish Somaliland from a generic desert.

The first movement in the Prompt-to-Place Method is therefore a shift from subject to setting. “Somaliland” becomes a road outside Hargeisa, a coastal view near Berbera, or a highland landscape near Sheikh. “Traditional culture” becomes a particular object, fabric, building, or activity. “Warm colors” becomes late-afternoon light falling across pale stone and red-brown earth.

This is not simply a matter of adding more words. AI image systems respond to relationships between words. A prompt combining “coastal Somaliland,” “limestone cliffs,” “Indian Ocean light,” and “a quiet fishing settlement” produces a different visual direction from one combining “dry plateau,” “acacia trees,” “camel caravan,” and “long evening shadows.” Each phrase narrows the field of possible images.

The creator also chooses what kind of artwork the prompt should become. A documentary-style image suggests recognizable textures and ordinary light. A painted interpretation may allow stronger color and visible brushwork. A graphic poster can simplify people and buildings into shapes. The place remains a central reference, but the chosen style changes what the viewer notices.

The difference resembles the difference between a map and a painting. A map emphasizes position and relation. A painting may emphasize atmosphere, memory, or feeling. Both can refer to the same landscape, yet neither contains the whole place.

Why Somaliland Cannot Be Reduced to a Desert

Outside observers have often represented the Horn of Africa through a narrow visual vocabulary: sand, drought, livestock, and distant horizons. These elements are real, but they are incomplete. Somaliland contains dry plains, mountain ranges, rocky coastlines, urban streets, pastoral routes, and settlements shaped by trade and migration. A prompt that uses only “desert” may produce an image that looks geographically plausible while remaining culturally vague.

The Golis Mountains, running along much of the northern Somali region, offer a useful correction to that flattened picture. Their ridges and escarpments create changes in elevation, shadow, vegetation, and distance. Near the coast, the relationship between mountain and sea can become part of the composition. Inland, the same broad region may appear harsher, more open, and more strongly marked by seasonal conditions.

Somaliland’s cities add another layer. Hargeisa is not simply a backdrop for traditional imagery. It is a living urban center with roads, shops, painted walls, transport, construction, and public spaces. Berbera carries the visual presence of a historic port, while smaller settlements connect urban life to pastoral and coastal economies. When these distinctions enter a prompt, the generated artwork has more than an African setting; it begins to have a particular geography.

Cultural references work in a similar way. The Somali language, the colors and patterns found in local textiles, the importance of poetry, the presence of camels, and the forms of hospitality associated with Somali life all carry meaning beyond decoration. Yet an AI system may not understand those meanings as a person from Somaliland would. It may treat a cultural object as a visual token rather than as part of a social world.

A single word can therefore carry two different kinds of information. To a person, “camel” may suggest livelihood, movement, status, history, and landscape. To an image model, it may be linked mainly to a collection of shapes seen in captions and pictures. The creator’s work lies partly in closing that gap.

What the Machine Knows - and What It Does Not

AI image generators create images by recognizing patterns in their training data and producing a new arrangement that matches a written request. They do not retrieve one hidden photograph of Somaliland and place it on the screen. Instead, they calculate which visual features are likely to belong together: a type of light, a landscape, an architectural form, a garment, or an artistic style.

This process explains both the power and the uncertainty of generated artwork. The system can combine a coastal horizon with a painted texture, or place a camel caravan beneath a dramatic sky, even if no single source image contained that exact arrangement. At the same time, it may confuse regional details, invent architectural features, or produce clothing that looks broadly “traditional” without belonging to a specific community.

The result can be visually convincing and culturally imprecise at once.

That tension becomes visible when a creator reviews the first generation. Perhaps the mountains are too tall, the road is too clean, the colors are overly saturated, or the people’s clothing does not fit the intended setting. Perhaps the image has captured a mood but lost the place. Revision then becomes a form of cultural editing. The creator removes some signals, strengthens others, and asks the system to move closer to the original idea.

A prompt might evolve from “cinematic Somaliland landscape” to “a dry road near Hargeisa at sunset, low acacia trees, distant rocky hills, restrained earth colors, documentary composition, no generic desert dunes.” The later wording does more than improve appearance. It rejects an image of Somaliland assembled from familiar but inaccurate shortcuts.

There is a surprising connection here: the more specific the prompt becomes, the less the creator is asking the machine to invent freely. Specificity does not eliminate imagination. It relocates imagination from the final image to the choices that shape it - what to name, what to leave out, and what visual history to invite.

That matters because AI artwork is often described as automatic. In practice, place-based work depends on a chain of human judgments. The creator selects the location, identifies meaningful details, notices errors, compares versions, and decides when an image has moved from a generic regional picture toward a recognizable interpretation of Somaliland.

A Creator, a Place, and the Work of Revision

The clearest way to understand this process is through the kind of artwork associated with Labojohar Arts: an image can begin with a broad cultural idea and gradually acquire local character through repeated attention. A creator may start with a scene connected to Somaliland’s landscape, then adjust the prompt after noticing that the first result has the wrong balance of land and sky. A second version may place too much emphasis on spectacle. A third may bring the viewer closer to the road, the textile, the building, or the person at the center of the composition.

The final image carries traces of those decisions even when viewers never see the earlier versions. A subdued palette may reflect a choice to avoid exaggerated “exotic” color. A wide horizon may come from the openness of the landscape rather than from a generic preference for cinematic scale. A carefully placed sign or pattern may connect the image to Somaliland more firmly than a dramatic mountain range could.

This is where the creator’s relationship to place becomes visible. Someone familiar with Somaliland may recognize that a generated coastline feels wrong because the shore has been made too tropical, or that an urban scene has been stripped of the density and dust of an actual street. Someone working from memory may use different clues: the color of evening light, the shape of a ridge, the feeling of distance between settlements, or the visual presence of camels moving across open ground.

An AI image cannot confirm that it has represented Somaliland faithfully. It can only respond to the material placed before it. The responsibility for meaning remains with the person who frames the request and evaluates the result.

The artwork’s human story is therefore not hidden behind the technology. It is distributed across the image: in the choice of place, the cultural references, the rejected errors, and the final balance between recognition and interpretation. Labojohar Arts becomes part of that story not because AI replaces artistic judgment, but because the technology gives judgment a new surface on which to work.

The Place That Appears Between Memory and Machine

A generated image of Somaliland is never only a picture of land. It is also a record of what the creator considers worth preserving in visual form. The prompt may mention mountains, roads, textiles, camels, architecture, or sea light, but beneath those details is a decision about which parts of place should remain visible when translated into pixels.

That translation reveals something broader about human culture. People have always used tools to turn memory and geography into images - stone, ink, paint, film, and now computational systems. Each tool changes the result, but none removes the need to choose what a place means.

An AI image may therefore be most interesting where it fails to settle the question. Its mountains may be slightly unfamiliar, its light partly invented, and its cultural details unevenly understood. Yet in that imperfect space, a new kind of artwork appears: Somaliland as encountered by a human imagination, reconstructed through a machine, and left open for the viewer to recognize, question, and wonder about.

End of chapter one. 4 more chapters and about 8,550 words in the full book.