Gemini Gets Deeper with Research and Reasoning

Google's Gemini chatbot platform now boasts "Deep Research," a feature enabling it to analyze research problems and produce comprehensive reports. This advanced tool leverages "advanced reasoning" and "long context capabilities" to generate research briefs within the Gemini apps, exportable to Google Docs.

Deep Dive into Deep Research

Currently exclusive to Gemini Advanced, part of the Google One AI Premium Plan ($20/month), Deep Research acts as a research assistant, analyzing web information relevant to user queries. The results are presented as summaries with links to original sources.

  1. Users pose a question.
  2. Deep Research creates a research plan for review.
  3. Upon approval, Deep Research refines its analysis through multiple search iterations.
  4. Finally, a report of key findings is generated.

Initially available in English on desktop and mobile web, Deep Research can be accessed via the "Gemini 1.5 Pro with Deep Research" option. Mobile app integration is expected in early 2025. This feature aims to save users hours of research time, as explained by David Citron, Gemini's product director. See more about Gemini 2.0.

Ethical Considerations and Potential Impacts

Deep Research raises ethical concerns, particularly in education. Over-reliance on AI for research could hinder critical thinking skills, as highlighted by Jessica Grose in The New York Times. A study even linked heavy ChatGPT use among students to procrastination, memory loss, and lower GPAs. Furthermore, Deep Research's web scraping practices may impact publishers' ad revenue, similar to the effects of AI Overviews in Google Search. While Google emphasizes collaboration and respects paywalls, publishers face a difficult choice between allowing scraping and maintaining search visibility. The long-term impact on web traffic remains to be seen. For more on AI's impact, see Google's New AI Agents.

Gemini 2.0 Flash: A Glimpse into the Future

Beyond Deep Research, Gemini users now have access to an experimental version of Gemini 2.0 Flash, Google's latest AI model. Available on desktop and mobile web, this experimental version promises improved performance and faster responses. The full version is slated for January 2025. However, some Gemini features may not be fully compatible during this experimental phase. Learn more about Gemini Advanced.