Google is set to enhance its Gemini Android app with the introduction of the Deep Research tool. Previously, this artificial intelligence (AI) feature was exclusively accessible through the web version of the platform. Launched in December 2024, the tool serves as a comprehensive research assistant, capable of devising multi-step research plans, conducting web searches, and compiling detailed reports on intricate subjects. Currently, this functionality is limited to paid subscribers of the Gemini service.
Deep Research AI Agent Being Added to Gemini App for Android
The Gemini team announced the rollout of the AI agent via a post on X (formerly Twitter). The rollout is currently in progress, meaning that it may take several weeks for users around the world to access the new feature on their Android devices. Once integrated, the Deep Research tool will be available under the Gemini Advanced drop-down menu.
Upon its initial release, Google indicated that the tool would support 45 languages, including Arabic, Bengali, English, French, Japanese, Russian, Tamil, and Vietnamese. Deep Research is driven by the newly launched Gemini 1.5 Pro, enhancing its ability to function as an effective research assistant.
When a user inputs a prompt regarding a complex issue, the AI agent formulates a multi-step research plan. Users are given the option to review and modify this plan, allowing them to add, remove, or change list items. Subsequently, the AI examines relevant academic articles, research papers, and other credible sources to gather pertinent information. It also investigates the latest developments in the field, analyzes current trends, and considers future implications.
Moreover, if necessary, the agent can perform additional web searches to obtain context on related topics. After accumulating sufficient information, it produces a comprehensive report on the topic at hand. Throughout the entire process, users retain the ability to intervene and adjust the research plan as they see fit.
As it learns from the retrieved data, the agent can initiate further web searches for a deeper understanding of the subject. Once the agent compiles enough insights, it generates a detailed report. Users can intervene at any stage, providing the flexibility to edit the research plan or add follow-up prompts to enhance the output.
