Is it just me, or is Gemini’s image editing going down the shitter FAST?

Exploring the Decline in Gemini’s Image Editing Capabilities: A Recent User Perspective

In recent weeks, I’ve been experimenting with Gemini’s image editing features, transitioning from my previous experience with ChatGPT’s image generation. Initially, Gemini impressed me with its ability to produce and modify images based on simple prompts, often surpassing my expectations. However, in the past few days, I have observed a noticeable decline in its performance, raising concerns about its current reliability.

When I first started using Gemini, it demonstrated remarkable responsiveness and accuracy. For instance, requesting it to add my brothers in suits to a photograph was straightforward—without needing to specify that their faces remain unaltered, it understood implicitly. The results were impressive, with the software correctly placing them in full suits, maintaining facial integrity.

Recently, however, I’ve encountered issues that suggest a regression in functionality. Now, I find myself having to explicitly instruct Gemini to leave faces untouched, and even then, it often produces incomplete or unnatural results, such as only adding a jacket rather than a complete suit. It appears to struggle with fulfilling complex, multi-faceted editing requests unless it compromises other aspects of the image. Additionally, after making a few edits to rectify these issues, the system abruptly informs me that I have exhausted my available edits—underscoring a frustrating user experience.

This rapid shift from promising to problematic use cases feels like a step backward in terms of AI-assisted image editing. For users who rely heavily on such tools for creative or professional purposes, this decline can be particularly disappointing. It raises questions about the stability and consistency of Gemini’s image editing capabilities moving forward.

Have others experienced a similar drop in quality or reliability? As AI-driven tools continue to evolve rapidly, maintaining consistent performance is crucial for user trust and utility. I look forward to seeing how Gemini addresses these challenges in upcoming updates, but for now, the recent decline has certainly tempered my enthusiasm.

Conclusion

While Gemini initially showed significant promise as an AI image editing tool, recent performance issues suggest a need for developers to refine and stabilize its features. Continued user feedback and constructive critique are essential to guide improvements and ensure these tools meet the high standards users expect.