I compared ChatGPT and Gemini's AI image generation - and a single prompt tweak made a big difference

2026-06-06 · Source: News and Advice on the World's Latest Innovations | ZDNET · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Novice, short

Summary

ZDNET contributor Kerry Wan and Tharon Green detail a prompting technique suggested by Lance Whitney, where users ask an AI chatbot to generate a detailed image prompt for its corresponding image generator. This method, demonstrated with a request for a "sunflower made of sheet metal in a pencil drawing style," was tested using Gemini and ChatGPT. Gemini's Nano Banana produced a "decent image" from its generated prompt, which included specific details like "sculpted pieces of metal sheet" and "visible welds." ChatGPT's image generator yielded a "very successful image" from its prompt, specifying elements such as "thin, slightly curved metal plates with visible seams, rivets, and subtle bends." This approach helps users create more precise and higher-quality images by leveraging the chatbot's ability to craft optimized queries and avoid language that might be flagged.

Key takeaway

For marketing or design professionals struggling with AI image generation, you should adopt a meta-prompting strategy. Instead of directly crafting complex prompts, ask your preferred chatbot (like Gemini or ChatGPT) to generate the detailed image query for you. This approach leverages the AI's understanding of its own generator's nuances, leading to more precise and higher-quality visual outputs, while also circumventing potential content flags. Experiment with this technique to streamline your creative workflow and achieve better results.

Key insights

Asking an AI chatbot to generate its own image prompt significantly enhances output quality and detail.

Principles

• Chatbots can optimize prompts for their own generators.
• Detailed prompts yield better AI image results.
• AI-generated prompts avoid flagged language.

Method

Provide a basic image concept to a chatbot, then instruct it to generate a detailed, optimized prompt for its corresponding image generator. If too long, request a shorter version.

In practice

• Use Gemini or ChatGPT to craft image prompts.
• Specify material, style, and background elements.
• Request shorter prompts if initial output is excessive.

Topics

• AI Image Generation
• Prompt Engineering
• ChatGPT Images
• Gemini Nano Banana
• Generative AI
• AI Workflow Optimization

Best for: Prompt Engineer, AI Student, Marketing Professional

Related on AIssential

• ChatGPT's Nano Banana — OpenAI's ChatGPT Images 2.0 significantly improves text and realism in image generation, expanding AI applications.
• Gemini 3 Deep Think — Gemini 3 Deep Think excels at generating complex, detailed SVG images from intricate prompts.
• This One Prompt Unlocks ChatGPT Images 2.0 — Reverse-engineering image prompts with ChatGPT allows users to recreate visuals and learn effective prompting techniques.
• 7 Ways to use ChaGPT — AI tools like ChatGPT offer diverse applications beyond search, enhancing productivity and personal tasks.
• Why I Ditched ChatGPT: A One-Month Showdown Between the Big Three AI Pro Plans — ChatGPT's undisputed leadership in generative AI is now challenged by competitive premium alternatives like Gemini Pro and Claude Pro.
• Developer Guide: Nano Banana 2 with the Gemini Interactions API — Nano Banana 2 enhances image generation with real-world grounding via Google Search and subject consistency from reference images.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by News and Advice on the World's Latest Innovations | ZDNET.