Hi Gamma Team,
I have noticed a significant drop in the quality and relevance of the AI-generated images in Gamma recently.
Previously, the visuals were consistently accurate, professional, and well-aligned with the presentation context. Lately, the output has become largely unusable due to several recurring issues:
Irrelevant Context: The generated images frequently miss the topic and intent of the slide content.
Visual Artifacts & Errors: Generated assets show noticeable rendering flaws, distortions, and composition errors.
Exaggerated Facial Expressions: Whenever people are depicted, they often show unnatural, extreme emotions rather than professional, balanced expressions.
This currently adds extra manual work to replace nearly every visual asset. Please review the underlying image generation models or default prompting logic to bring the output back to the high standard we are used to.
Best regards,
Philipp