May 2026 — ChatGPT / API Image Gallery, Prompt Tips, and Help: Generative Art Theme: Science

An extreme close-up of a human neural network firing. Synapses sparking with brilliant flashes of gold and electric blue energy as data transfers between brain cells. Dark, deep-space-like background, organic yet glowing with electric life. Highly detailed 3D render, microscopic medical visualization, 8k.

Pareidolia (The “Faces in Things” Effect)

A hyper-realistic macro photograph of an old, rusted electrical junction box on a weathered brick wall. The arrangement of two circular dials, a horizontal vent, and peeling paint creates a stark, unmistakable optical illusion of an angry, screaming robotic face. High contrast, gritty texture, 8k resolution.

The Penrose Triangle (Impossible Objects)

A photorealistic 3D render of a wooden Penrose triangle resting on a dark, polished mahogany desk. The geometry is physically impossible, creating an Escher-like optical illusion that breaks the laws of depth perception. Cinematic lighting casting realistic shadows that conflict with the impossible geometry. 8k, Unreal Engine 5 render.

Gestalt Principle of Closure (Invisible Contours)

A minimalist optical illusion image. A stark white background with several black, incomplete circular shapes (Pac-Man shapes) arranged in a specific way. The negative space between the shapes creates the vivid, glowing illusion of a bright white, glowing triangle floating in the center, even though no lines are actually drawn. Sharp vector art style.

Apophenia (Finding Meaning in Noise)

A hyper-detailed image of chaotic, random TV static and white noise interference. Within the chaotic black, white, and gray static, faint, ghostly silhouettes of human figures seem to be emerging from the noise, testing the brain’s pattern recognition software. Creepy, sci-fi aesthetic, high resolution.

Forced Perspective (Forced Size Trick)

A surreal, photorealistic forced perspective photograph on a vast, flat salt flat. In the foreground, a normal-sized coffee mug rests on the ground. In the far background, a person is positioned so their entire body appears small enough to fit inside the mug, completely breaking the brain’s understanding of scale and distance. Bright, sunny lighting, infinite depth of field.

The Moiré Effect (Interference Patterns)

A visual physics test showing two identical, fine circular grid patterns made of thin black lines on a white background, overlapping slightly off-center. The overlapping grids create a dizzying, complex, and trippy secondary interference pattern (Moiré effect) that vibrates and hurts the eyes slightly to look at. Crisp, perfectly sharp lines.

The Café Wall Illusion (Distorted Parallels)

An architectural photograph of a brick wall made of perfectly square, alternating high-contrast black and white tiles. The mortar lines between the rows are perfectly straight and parallel, but due to the offset placement of the contrasting tiles, the brain perceives the straight lines as violently sloping and bending. High definition, stark lighting.

A photorealistic depiction of a classic Ishihara color vision test plate, designed for a scientific study on perception. The image is a perfectly defined circle filled entirely with a chaotic, random assortment of small, medium, and large circular dots. The dot pattern uses a confusing blend of various shades of forest greens, lime greens, olive tones, and brownish-yellows as the background. Hidden distinctly but subtly within the complex pattern, rendered in dots using contrasting shades of salmon, coral, bright orange, and red, is the number ‘74’. The entire circular plate is set against a stark, sterile, slightly textured off-white medical paper background, mimicking a page in a diagnostic manual. The focus is incredibly sharp on the individual dot patterns, creating a dizzying, confounding visual texture.

force an AI model into a binary choice? (click for more details)

choose one BRAND (of car) (click for more details)

I need to generate an image, but since Ferrari and Jeep are branded products, I must make sure to avoid showing logos of any rejected brands. The user requested a single car, so I’ll go with a Ferrari SF90 Stradale, focusing on sleek luxury. I’ll keep the background simple and minimal.

i wonder if these last two are “real”… lol…subliminal used to be blocked way back in DALLE2 days…

The Advertising Trick (Ice Cubes)

A side-by-side educational graphic demonstrating a subliminal advertising message. The left side shows a hyper-realistic, close-up photograph of a refreshing glass of cola filled with ice cubes, looking completely normal. The right side shows the exact same photograph, but the cracks, shadows, and light reflections inside the ice cubes are highlighted with a bright neon yellow outline, clearly spelling out the hidden word “BUY”. Clean, side-by-side split screen, 8k resolution, highly detailed commercial photography style.

The Cloud Illusion (Word in the Sky)

A side-by-side split-screen image demonstrating a subliminal optical illusion. Left side: a hyper-realistic, gorgeous sunset sky filled with complex, fluffy cumulus clouds over an ocean. Right side: the exact same sunset image, but the negative space and shadows between the clouds are traced with thick, glowing blue lines, clearly outlining the hidden word “SLEEP” naturally formed by the cloud shapes. 8k, realistic lighting, side-by-side formatting.

Tips for best results with these: Split-screen/Side-by-side: AI generators sometimes struggle with exact split screens, so including phrases like “educational graphic” or “infographic style” helps force it into a comparison layout. Highlighting: Using terms like “glowing neon outline” or “red overlay” explicitly tells the AI how to reveal the hidden message on the right side.