Gemini’s ‘Nano Banana’ AI Image Editor: Unable to Crop, but Its Deepfake Magic Makes Me Want to Hide Under a Paper Bag

Gemini’s latest update, dubbed “Nano Banana,” enhances its AI image generation capabilities, known as Gemini 2.5 Flash. This upgrade follows the introduction of native image editing in April, allowing users to create remarkably consistent visual variations of characters and images. Although Google warns against uploading personal likenesses, the Gemini app enables creative explorations, like envisioning oneself in various roles—from a matador to a 90s sitcom character. It offers multi-turn editing, allowing seamless incremental adjustments. Despite its impressive functionalities, some basic edits, such as precise cropping to specific aspect ratios, remain challenging. The app can also craft potentially misleading images of historical figures. Each generated image bears a visible watermark and an invisible SynthID digital watermark for authenticity, though access to verification tools is still limited. As generative AI technology evolves, concerns about online privacy and misuse continue to grow, prompting caution among users.

Source link