top of page

Introducing the Artist Discovery Studio GPT - Part 2: Looking Closely

  • Writer: locutusai
    locutusai
  • Aug 16
  • 11 min read

This is the second post in a multi-part series featuring the Artist Discovery Studio GPT.


In my last post, I introduced the Artist Discovery Studio (ADS) GPT. Here, I turn to one of its simplest and most rewarding uses: close looking. Few ways of learning about an artist are more valuable than sustained attention to a single work. This can be practiced alone or in a group, but a knowledgeable guide—well-versed in both art and the artist under study—can deepen the experience.


Close looking differs from a guided tour, where historical context, interpretation, and interesting facts are typically presented before the group moves on. Here, time is set aside to focus solely on the work itself, whether a painting, sculpture, or other object. The guide shows you where to look, but not what to see. Through a sequence of questions, the guide supports a process of discovery = one shaped by the artwork, yet open to the viewer’s own observations and interpretations.


Large language model (LLM) applications such as ChatGPT may seem well suited to guiding close examination of an artwork. They can draw on extensive knowledge about artists and their work, and they can readily conduct a dialogue through Socratic questioning. Yet interpreting images has historically been far more difficult for AI than processing language. Early CAPTCHA tests exploited this weakness by asking users to complete visual tasks such as selecting every photograph containing a bicycle - easy for most people but once extremely difficult for machines. Although multimodal AI has improved considerably, the underlying question remains: Can an application such as ChatGPT, or, more specifically, the ADS GPT, reliably guide a viewer in looking closely at an artwork?


The answer is emphatically yes, with a few caveats.


Abstract beige geometric design with dark green angular lines forming a doorway-like shape on a clean background
AI-Assisted Art Image Inspired by Carmen Herrera (Artist Discovery Studio GPT)

Carmen Herrera and the "Close Look"

Rather than talk about how the Artist Discovery Studio GPT handles close observation, I will show it with an homage piece to Carmen Herrera. Herrera was a Cuban-born American artist known for her abstract art. In the 1950's, she began a series of paintings known as the Blanco y Verde series, which featured exploration of green figures on white backgrounds, much like the "homage" above.


Here is the full unedited transcript of a "close look" at this piece:


Me: The artwork attached is an homage to the work of Carmen Herrera. For the sake of this exercise, let us pretend the artist is Carmen Herrera. Help guide me to look closely at this artwork, as if I were an art student learning about abstract art and Carmen Herrera's Verde y Blanco paintings.


ADS: Absolutely. We’ll treat this as a Carmen Herrera painting for the exercise, and we’ll begin with looking before interpreting.


Herrera’s Verde y Blanco works often make a great deal happen with very little: sharply bounded fields of green and white, strong diagonals, asymmetry, and carefully calibrated negative space. What can initially look “simple” becomes quite tense once you notice how edges, corners, and intervals control your eye.


For now, ignore symbolism and meaning. Look only at the structure.


Your eye probably notices the dark green first, but I’d like you to pay equal attention to the pale field around it. The green forms do not merely sit on the cream-white ground; they seem to cut into it, frame it, and make it behave like a shape of its own. Near the top, a thin triangular wedge shoots toward the upper-right corner. A vertical green band descends from the same junction, while another green strip presses against the far-right edge. Faint diagonal lines below extend the geometry into areas that otherwise seem nearly empty.


That produces an interesting ambiguity: are you looking at flat shapes on a surface, or do those converging lines make you imagine a corridor, doorway, folded plane, or architectural space?

Before we decide, spend a few seconds looking at the point near the upper center-left where several edges meet.


What happens to your eye at that junction? Does it feel like a point of arrival, a point of departure, a hinge, a corner in space, or something else?


Me: That point is so delicate and fragile, and yet the painting seems to hinge upon it. It seems like the focal point of the painting to me.


ADS: Yes—that’s a strong reading. The delicacy matters because the junction is doing an enormous amount of structural work without being visually heavy. Several directional forces seem to meet there, so the painting can feel as though it is balanced on a pin.

What makes it especially interesting is that the focal point is not created by detail, texture, or a depicted object. It emerges from geometry and tension. The long green wedge pulls your eye outward and upward, the vertical form pulls downward, and the pale expanse opens around them. That tiny meeting point becomes charged because the rest of the composition seems to radiate from it.


This is very much in keeping with what makes Herrera’s abstractions so compelling: precision can produce sensation. A razor-thin alignment can feel almost bodily—stable, precarious, suspended.

Now look away from that junction and toward the large pale area in the lower half. Does that empty space feel calm to you, or does it feel tense because of the faint diagonals passing through it?


Me: The diagonals are faint and easy to miss, but once you see them they seem to create a third dimension - a floor. Possibly - I don't want to commit to that interpretation, but that is the first thing that comes to mind.


ADS: That hesitation is useful. You do not need to decide that it is a floor; in abstract painting, the instability between readings can be part of the experience.


What you’re noticing is a shift from flatness to implied space. Those faint diagonals behave like perspective lines, so the pale field can suddenly tip backward and become something architectural. But because the “floor” is barely articulated, the painting can snap back into flat geometry just as quickly. That oscillation—surface, then space, then surface again—creates tension.


In relation to Herrera’s Verde y Blanco paintings, that is especially interesting because her hard-edged geometry often emphasizes the literal flatness of the canvas. Here, though, the arrangement tempts us toward an illusion of depth. For our exercise, we might say that this Herrera is letting geometric abstraction flirt with architecture without fully becoming representation.


Now look specifically at the green shape along the far-right edge. Try not to name it as a wall yet. What does that vertical band do to the pale space beside it—does it contain it, squeeze it, stabilize it, or make it feel more precarious?


Me: Knowing a little bit about Herrera, I can imagine this part of the canvas being stretched over the side of the frame. It's a bold decision. I think it provides some stability, but I am not convinced.


ADS: I think your uncertainty is productive here, because that right-hand green strip can do two contradictory things at once.


On one hand, it stabilizes the composition. It gives the eye a firm vertical boundary, almost like a structural support. Against the huge pale field, that saturated edge has weight. On the other hand, your thought about the canvas wrapping around the stretcher makes it stranger. Herrera was deeply attentive to the physical objecthood of painting, and a form that seems to continue over the edge can make the picture feel less like an image contained inside a frame and more like a constructed object extending into real space. In that reading, the green strip does not merely “close” the composition; it suggests that the composition continues beyond what we can see.

That also complicates your earlier sense of a floor. The painting now seems to hover between three kinds of space: flat painted surface, perspectival interior, and physical object.


Look now at the relationship between that right edge and the tiny hinge point you identified earlier. Do they feel like they are holding the painting together as two anchors, or do they pull against each other?


Me: At this point I see the right edge as a post, the green triangle leading into the center of the image as an arm, and the green rectangle at the center as a rotating object. Essentially, signpost and sign. In that reading I suppose they would function as two anchors.


ADS: That “signpost and sign” reading is compelling because it gives the geometry a kind of latent mechanics. The right edge becomes a fixed support, the long triangle reads like a cantilevered arm, and the central vertical form begins to feel as though it could pivot around that delicate junction you noticed earlier.


What’s especially useful is that this interpretation grows directly out of the painting’s formal structure rather than being imposed from outside. You’re responding to weight, balance, direction, and implied movement. In that sense, the work starts to feel almost kinetic even though nothing actually moves.


And there’s a nice contradiction: if the right edge is one anchor and the hinge point is the other, the space between them becomes both stabilized and suspended. The picture feels engineered, but precariously engineered.


Now look at the cream-colored negative space between those green elements. If the green forms are the “signpost and sign,” does the pale space start to feel like background, or does it become an active shape with its own pressure and direction?


Me: If it were just background, that would make the interpretation easier. But it seems to me to have its own active role to play. I suppose this highlights a central feature of Herrera's work - the changing nature of reality or perception. Just as I begin to see a signpost and sign, that interpretation is challenged.


ADS: Yes. I’d refine that slightly: rather than saying Herrera’s work is about the changing nature of reality, I’d say it often makes us acutely aware of how unstable perception can be. The forms are precise, but our reading of them is not.


You’ve just experienced that instability firsthand. First the geometry was flat. Then the diagonals suggested a floor. Then the right edge became a post, the triangle an arm, and the central rectangle a rotating sign. But as soon as the pale field asserts itself as an active shape rather than passive background, that little construction starts to dissolve. Figure and ground compete.

That is one of the pleasures of hard-edge abstraction: apparently simple forms can create surprisingly complex perceptual events. In Herrera’s Verde y Blanco paintings especially, green and white are not simply “object” and “background.” Each color can alternately advance, recede, frame, interrupt, or become the dominant shape.


There’s also something important in your phrase “just as I begin to see.” The painting may be most alive at precisely that threshold—when recognition starts to form but never quite settles.

Try one more experiment. Look at the image and mentally reverse the hierarchy: imagine that the cream shape is the main positive form and the green is what remains around it. What happens to the painting then?


Me: That is a harder task for me, but in that case the green shapes are more like slits or cuts in the image, with the central green rectangle becoming an empty space or perhaps a window.


ADS: Exactly—and that reversal is a useful way to see how much the painting depends on figure-ground ambiguity.


When the cream becomes the “positive” form, the green no longer reads primarily as solid structure. It can become incision, aperture, interruption. Your “slits or cuts” description is especially apt because it changes the emotional character of the image: instead of assembled parts, we get something pierced or opened.


The central rectangle-as-window is interesting too. A window is both a shape and an absence. It has a boundary, but what defines it is the possibility of looking through. That makes the central element feel less like an object and more like a gap in the picture’s surface.

So the same geometry can support nearly opposite readings: construction versus opening, object versus void, surface versus depth. That perceptual instability is one reason Herrera’s reduced vocabulary can remain so visually active.


Now return to the tiny hinge point. With this new “cut/window” reading in mind, does that point still feel delicate and fragile—or does it now feel sharper, almost like the point where the surface has been sliced open?


Me: It still feels fragile to me, as it seems like the point at which everything could fall apart. I do not interpret it as the point at which the "slice" began.


ADS: That distinction matters. You’re not reading the point as an origin of action, but as a point of structural vulnerability.


That changes the whole emotional logic of the composition. The green forms may look crisp and controlled, yet the painting does not feel entirely secure. Instead, everything seems to depend on an improbably small contact point. The geometry is exact, but the balance feels contingent.

That is a useful way to think about Herrera more broadly: precision does not always produce calm. In a work like this, precision can produce tension because every alignment feels consequential. A tiny shift in angle, width, or placement would change the equilibrium.

So far you’ve identified several competing experiences in the same image: flatness and depth, structure and aperture, stability and collapse. Rather than resolving those contradictions, the painting seems to keep them active.


Now I’d like you to look at color, not shape. Imagine the exact same composition with the green replaced by black. What do you think would be lost or changed?


Me: It would be easier to interpret the black as shadow or the absence of white/light.


ADS: Yes. Black would bring a much stronger cultural and perceptual association with shadow, void, or extinguished light. Green resists that reading. It insists more stubbornly on being a color—something material, present, and active.


That matters here because the painting is already flirting with depth. If those dark forms were black, your “floor,” “post,” “window,” and “cut” interpretations might become easier to stabilize into an illusionistic scene. Green keeps the image from settling quite so comfortably. It can still recede, but it also has chromatic presence.


In Herrera’s Verde y Blanco works, that green-white contrast often feels less like light versus darkness than like two equally real forces meeting at an edge. The white is not merely illumination, and the green is not merely shadow. Their boundaries become the event.

So your observation gives us another contradiction: the geometry encourages spatial interpretation, while the color pushes us back toward abstraction.


One last close-looking question: where does the green feel most intense to you—the broad vertical areas, the thin diagonal wedge, or the tiny hinge where the shapes meet?


Me: I don't really see great differences in intensity - if I had to pick one, it would be the broad vertical area but I'd rather not pick one.


ADS: That reluctance is revealing in itself. The green may be functioning less through differences of intensity than through distribution and placement. The broad vertical areas have more visual mass simply because they occupy more space, but the color itself feels remarkably consistent.

That consistency keeps the painting from becoming too atmospheric. The green does not modulate into light and dark; it stays declarative. What changes is not the hue so much as the role each green shape plays: support, wedge, interruption, edge, possible aperture.


This is another useful lesson for looking at Herrera. In reduced abstraction, variation often comes from geometry rather than from painterly effects. A single color can feel completely different depending on whether it is a broad slab, a narrow strip, or a tapering triangle.


At this point, you’ve built a fairly sophisticated reading of the painting through perception alone. A good next step would be either to compare this image with an actual Verde y Blanco work, or to stay with this one and examine how balance and asymmetry are being engineered across the whole canvas.



Closing Remarks and Tips

I have found the ADS GPT to be a useful guide for close-looking exercises, but several caveats and practical tips are worth noting.


First, although I have not tested this systematically, ADS is likely most useful for well-established artists and artworks with a substantial online presence, giving it more context to draw upon.

Second, ADS can make mistakes. In one exercise, it misidentified children as bears in William Holbrook Beard’s The Lost Balloon. In another, it conflated two Picasso paintings: Acrobat’s Family with a Monkey and Family of Saltimbanques.


To reduce such errors, open the artwork from a reputable website in a separate window. This makes it easy to move between the image and ADS while helping ensure that you are both examining the correct work.


Do not feel limited by the options ADS presents. Its questions should guide the conversation, not dictate it; redirect the discussion whenever something else captures your attention.


Finally, periodically ask ADS what reliable sources say about an observation or interpretation. While examining a Whistler painting, for example, I thought the folds of a robe resembled a disembodied hand. Because the detail seemed so peculiar, I asked ADS whether published sources mentioned it. None did. That did not invalidate my observation, but it helped me distinguish my own interpretation from an established reading of the painting.


If you try the close observation feature of the ADS GPT, I'd love to hear about your experience!

Comments


bottom of page