Proposal: Making ChatGPT Better than Gemini for Image Creation & Editing

Proposal: Making ChatGPT Better than Gemini for Image Creation & Editing

Who I am:
I am a paid ChatGPT user and professional oil painter. I use AI to support storytelling, cinematic scenes, and artwork. As an artist, I can spot details where image generation feels unnatural or incomplete.

Problem:
While ChatGPT’s DALL·E is strong in creativity, its editing ability is weaker than competitors (like Gemini). When correcting an image, instead of applying the exact change, the tool often alters unrelated details (face, jewelry, clothing, or background). This makes it unreliable for serious creative work.

Comparison:

Gemini and other tools follow correction prompts more strictly.

ChatGPT sometimes ignores or over-modifies prompts.

This difference gives other AI tools an upper hand in professional art workflows.

Suggestions to Improve:

  1. Strict Edit Mode → Only change what is explicitly requested. Lock other details.

  2. Character Locking → Preserve the same face/identity across multiple generations.

  3. Detail Fidelity → Ensure ornaments, fabric, skin textures, and costumes remain consistent.

  4. Pose & Lighting Control → Option to use reference poses or specify cinematic lighting (golden hour, candlelight, rain reflections).

  5. Reliability → Keep the image tool consistently available for paid users, with fewer shutdowns.

Artist Perspective:
As someone who paints with oil on canvas, I can contribute feedback on composition, light, realism, and texture. Even if I don’t know the technical side of mobile AI, I know what makes an image professionally correct.

Closing thought:
With these improvements, ChatGPT can surpass Gemini and others in image creation, giving artists a tool that combines the power of AI with the discipline of professional art.

Follow-Up Comment

As an oil painter, my perspective is more on composition, lighting, and realism. I may not know the technical side of AI, but I know what makes an image look professional and natural. That’s why I strongly feel edits should be precise — only changing what the user asks, without losing costumes, jewelry, or face identity. I hope my artist’s point of view can add value to improving this tool.

This topic was automatically closed after 19 hours. New replies are no longer allowed.