Feedback on Image Generation Behavior – Conversation Flow and Feedback Options

When generating images inside ChatGPT (e.g., using DALL-E), two issues disrupt the otherwise smooth and conversational user experience:

  1. Abrupt Conversation Cutoff After Image Generation:
  • After an image is generated, ChatGPT immediately ends the conversational flow.
  • This cutoff feels inconsistent compared to text-based interactions where a natural back-and-forth occurs.
  • Suggestion: Allow ChatGPT to briefly confirm the successful generation (e.g., “Here’s your image. Would you like any changes?”) to maintain conversational continuity without cluttering the experience.
  1. Missing Feedback Options After Image Responses:
  • After image generation, the toolbar only includes Copy, Download, and Read Aloud (how would ChatGPT “read aloud” an image?).
  • Thumbs-up and thumbs-down feedback buttons — present for text responses — are missing entirely.
  • Suggestion: Enable the ability to give direct thumbs-up/thumbs-down feedback on image generation behavior, just like on normal messages. This would allow more accurate, case-specific feedback on image quality and UX.

Why These Changes Matter:
Improving these behaviors would make image-based interactions feel more integrated with ChatGPT’s core conversational style and encourage better user feedback, helping continuously refine the image generation experience.