
Google has unveiled a new feature for its AI-driven chatbot, Gemini. This feature grants users precise control over generated responses, enabling them to modify specific text segments. Users can regenerate, adjust length, or delete selected portions. Currently, this feature is exclusive to the web app version and hasn’t been integrated into the app. Notably, Google has also introduced several new AI functionalities for Android devices.
In a recent post, Google revealed the introduction of a new feature for Gemini, stating that they are providing users with a more precise method to adjust responses. They explained that users can now select a specific portion of text in the Gemini web app, offer instructions, and receive an output closer to their desired response. The addition of this feature aims to enhance users’ creative control by allowing them to modify small sections of the response instead of regenerating it entirely. Users can pick and highlight any portion of the text after a prompt has been generated.
You can see a pencil icon in a circle by highlighting it right now. Besides the pencil, the Gemini logo is also visible. The text “Modify selected text” appears when the arrow is hovered over the icon. Four options are displayed when you click on this icon: Regenerate, Shorter, Longer, and Remove.
As the name implies, the options function. The text will be generated again after being modified using the regenerate function. The topic’s length can be altered or adjusted using Shorter and Longer, and it can be completely removed using Remove. Additionally, users have the option to include a prompt to change a particular section.
That is a useful addition. People who have worked with AI chatbots for some time know that often, an excellent response gets destroyed by just one sentence. Up until now, the only ways to solve it have been to completely redo the material or alter the prompt, then keep doing so until the intended outcome shows up. Users might find that making fine-grained adjustments using this function saves a significant amount of time.
- Claude Computer Use: Anthropic’s AI Can Now Control Your Computer - March 25, 2026
- Meta Avocado AI Model Delayed Until May: What It Means for the AI Race - March 16, 2026
- Cybersecurity Threats in 2026: Key Insights from the Cloudflare Threat Report - March 10, 2026



