The 'Deep Research' tool is available for Gemini Advanced subscribers.
Google has launched an AI tool that conducts research on your behalf and churns out detailed, easy-to-understand reports so you don't have to, according to a blog post from the company published on Wednesday.
The Gemini "Deep Research" tool runs on the Gemini 1.5 Pro model and is available on desktop and mobile via a web browser in English. To access it, go to www.gemini.google/advanced/, click the Gemini Advanced logo at the top left, and pick "1.5 Pro with Deep Research."
Once launched, you can ask Gemini to research a particular topic. In response, the tool will develop a multi-step research plan for you to revise or approve. If you approve it, the tool will start researching. It will browse the web for different takes and refine its analysis. This process will be repeated multiple times before a comprehensive report with key findings is delivered.
The report will be organized and include links to relevant sources. You can also ask Gemini follow-up questions or to make changes to the report. Once the final report is ready, it can easily be exported to Google Docs. According to Google, the entire process takes a few minutes but saves hours of research.
While this could help users save time, there's always the possibility that the AI hallucinates parts of its report, generating inaccuracies or falsehoods buried in a multi-page document. Or, the feature could be misused to complete school assignments where AI is not allowed, for example. The American Psychological Association, which oversees the academic APA Style Guide, recommends human oversight when using AI tools because it argues you have to take responsibility for that AI-generated work and an AI can't be considered an author.
If you're looking to try it out, though, Deep Research is only available to Gemini Advanced subscribers. Google is currently offering a free trial for the first month, but after that the subscription costs $20 a month in the US.
Separately, Google also announced Gemini 2.0 on Wednesday. The company hopes this new model will enable developers and testers to build more AI agents capable of accomplishing tasks autonomously. Deep Research, which essentially mimics a research assistant, is just a simple example of this vision (though it's not built on a 2.0 model).
The first model in the Gemini 2.0 family, an experimental version of Gemini 2.0 Flash, is available to all users on the desktop and mobile web -- with plans to launch it on the mobile app soon. Deep Research will also be added to the Gemini mobile app in early 2025.