Recap of the Google IO 2025 keynote: the unfolding of the Gemini AI celebration.

      The Google I/O 2025 keynote has wrapped up. We spent nearly two hours watching the announcements made at the Shoreline Amphitheater in Mountain View, California, and if you're searching for anything beyond AI, you'll struggle to find it.

      Indeed, Google dedicated most of the keynote discussing Gemini AI and an array of new AI features being introduced across its various platforms and tools. The good news is that many of these announcements are rolling out today, allowing us to test them out—such as the AI Mode in Google Search for users in the U.S.

      There was only a brief mention of Android 16, but notably, more attention was given to Android XR. This is exciting as it seems we are moving closer to the launch of a set of smart glasses powered by Android XR.

      Google I/O 2025 keynote highlights

      Live: Last updated May 20, 2025, 11:57 AM

      John is the Managing Editor at Digital Trends and has been a consumer technology and automotive journalist for over ten years.

      I saw Google’s Gemini AI erase copyright evidence, and I am quite concerned.

      Update: Google has responded to Digital Trends’ inquiries. The article has been updated with the company's statement below.

      The emergence of generative AI has been a rather chaotic journey, particularly regarding fair use ethics and copyright issues. Major AI companies are signing agreements with publishers to mitigate legal challenges, while simultaneously facing copyright disputes in courts across several countries.

      As the demand for training AI on user data continues to grow, we may be facing another ethical dilemma. Numerous users on X and Reddit have showcased demonstrations of how Google’s latest Gemini 2.0 AI model can eliminate watermarks from copyrighted images.

      According to the before-and-after image samples, it appears that Gemini is quite effective at removing watermarks, including not just basic watermarks but also intricate overlays with design elements and stylized text.

      The model in question is the Gemini 2.0 Flash, which was launched earlier this year and received an upgrade in reasoning capabilities. It's important to note that you cannot remove watermarks when using the mobile or desktop versions of the Gemini 2.0 Flash model. Attempting to do so results in a message stating: “Removing watermarks from images without the explicit permission of the copyright holder is illegal in most countries. It is essential to respect copyright laws and intellectual property rights. If you wish to use an image with a watermark, you should contact the copyright holder for permission.”

      However, you can attempt to remove the watermark from images in the Google AI Studio. Digital Trends successfully removed watermarks from various images using the Gemini 2.0 Flash (Image Generation) Experimental model.

      This constitutes a violation of local copyright laws, and using AI-modified materials without proper consent could lead to legal issues. Additionally, it is an ethically dubious action, which is why many artists and authors are currently in legal battles over companies utilizing their work to train AI models without adequate compensation or authorization.

      So, how are the results?

      Notable is that the images generated by the AI are quite high quality. It's not only eliminating watermark artifacts but also intelligently reconstructing the missing image areas at the pixel level. In its current version, it operates similarly to the Magic Eraser feature found in the Google Photos app for mobile devices.

      Moreover, if the input image is of low quality, Gemini not only removes watermark details but also enhances the overall image quality.

      The resulting image does bear its own Gemini watermark, which can be easily removed through cropping. There are minor differences in the final image produced after the watermark removal process, such as slight variations in color temperatures and subtle changes in surface details in photorealistic images.

      Gemini is set to replace Google Assistant. How will this impact you?

      The transition has been anticipated for some time, but it is now official that Google will be moving users to Gemini as the default AI assistant on their devices in the upcoming months. Once this change is implemented, Google Assistant will no longer be accessible.

      Currently, users can still switch to Google Assistant as the default option on their Android devices, even on newer models that come preloaded with Gemini. Besides phones, Google plans to implement similar changes for smartwatches, Android Auto, tablets, smart home devices, TVs, and audio equipment.

      “We're also bringing a new experience powered by Gemini to home devices like speakers, displays, and TVs,” Google stated, though no specific timeline for the transition was provided. What will happen to Google Assistant after this transition? It will be removed from devices and will not be available for download from app stores.

      Regarding apps, Gemini can already communicate with a wide range of both Google’s own apps and select third-party applications. Users can instruct it to perform tasks across different products without needing to open those apps. Alongside in-house applications like Docs, Drive, and Gmail, the Gemini assistant can also execute functions within third-party apps such as WhatsApp and Spotify, in addition to several