On August 27, 2025, 8:13 PM ET, Google began rolling out an upgrade to the Gemini app’s AI image editor that promises more reliable, realistic edits. According to Google, the refreshed tools can blend two photos, chain multiple edits in one session, apply the style of one image to objects in another, and even change outfits while preserving a person’s appearance.
Why It Matters
- Tackles a top user gripe: edits that subtly distort faces or pets now better preserve likeness.
- Expands creative control with multi-turn editing and cross-image “design mixing” for textures and patterns.
- Builds in provenance: images include visible labels and SynthID digital watermarks to signal AI assistance.
Details / Specs / Numbers
- Multi-turn editing: Iterate on the same image without unintentionally altering other regions.
- Photo blending: Combine multiple photos into a single scene.
- Design/style mixing: Transfer colors/patterns from one image to objects in another (e.g., apply butterfly-wing textures to clothing).
- Outfit changes with likeness preservation: Swap apparel or time-period styles while keeping facial identity consistent.
- Availability: Rolling out globally in the Gemini app for free and paid users; Workspace users also gain enhanced editing in the Gemini app experience.
- Safety & labeling: Visible “AI-generated” markers plus invisible SynthID watermarking.
Timeline & Official Statements
- April 30, 2025 — Google enabled in-app image uploads and basic editing in Gemini.
- August 26, 2025 — Google details the upgraded image model and capabilities; media reports highlight multi-step edits and improved likeness retention.
- August 27, 2025 — Follow-up coverage underscores global availability and feature depth (blending, design mixing, outfit changes).
Market/Industry Impact
Google’s move sharpens competition with OpenAI’s and Adobe’s AI-assisted image tools by emphasizing identity consistency and editable control over multiple steps. The inclusion of SynthID watermarking may help address misinformation risks that accompany more powerful editing features, though enforcement and cross-platform detection remain challenges. Analysts note the upgrade positions Gemini as a more full-featured mobile editor, potentially lifting user engagement across Android and iOS.
What to Watch Next
Expect iterative improvements to editing fidelity, broader device rollouts, and deeper ties into Google’s ecosystem (e.g., Photos, Workspace) as the company builds on Vertex AI’s underlying 2.5 Flash Image model. Also watch for expanded safety controls and clearer provenance standards across platforms. Google
TL;DR
- Gemini’s AI editor now supports multi-turn edits, photo blending, and design/style transfer.
- Likeness preservation improves for people and pets during outfit/style changes.
- All edited images include visible labels and SynthID watermarking.
FAQ
Q: Which devices can access the upgraded Gemini image editor?
A: Google says the features are rolling out in the Gemini app for both free and paid users on web and mobile, with Workspace users also supported.
Q: Can I blend two photos and keep my face looking accurate?
A: Yes. The update focuses on multi-image blending and better likeness preservation for people and pets.
Q: How are AI-edited images labeled?
A: The app adds a visible label and embeds Google’s SynthID watermark to indicate AI involvement.
External Sources
- Google Blog — https://blog.google/products/gemini/updated-image-editing-model/
- The Verge — https://www.theverge.com/news/765565/google-is-upgrading-the-gemini-apps-ai-image-editor
- TechCrunch — https://techcrunch.com/2025/08/26/google-geminis-ai-image-model-gets-a-bananas-upgrade/
- Google Cloud Docs (Vertex AI) — https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/image-editing
- Axios — https://www.axios.com/2025/08/26/nano-banana-google-ai-images
