Loss of Creative Precision in Image Generation – Request to Restore Previous Functionality

As a writer and educator who uses A-EYE to illustrate complex ideas—often satirical, historical, or psychological—I’ve noticed a sharp decline in image generation quality since the GPT-4o update. The visuals have become less accurate, less responsive to tone, and frequently disconnected from the prompts. A recent attempt to illustrate a satirical human history essay returned an image of a duck in a car windshield—entirely unrelated.

Before the update, the image tool could handle wit, irony, historical complexity, and layered storytelling. That capability was essential to my work, and to the broader community of creative users who rely on A-EYE not just for function, but for artistic fidelity.

I urge you to consider restoring access to the earlier image engine—or, better yet, offer a “precision mode” for those of us who use this platform for meaningful, mission-driven storytelling. This isn’t just about aesthetics—it’s about keeping the tool intelligent, responsive, and worthy of its original promise.

—Dr. Cima