Google has announced significant upgrades to its AI platform Gemini, introducing the “Deep Research” feature and the experimental rollout of the Gemini 2.0 Flash model. These updates enhance users’ research capabilities and improve the platform’s performance.
“We’ve built a new agentic system that uses Google’s expertise of finding relevant information on the web to direct Gemini’s browsing and research. Deep Research saves you hours of time, ” said David Citron, product director for the Gemini apps, for TechCrunch.
Deep Research
At the heart of the update is Deep Research, a feature that handles complex research tasks and delivers detailed reports. This tool, now exclusive to the Gemini Advanced tier (accessible via the $20/month Google One AI Premium Plan), utilises advanced reasoning and extended context capabilities to analyse data from across the web.
Deep Research enables users to submit a question, after which it generates a step-by-step research plan for their approval. Once approved, the system iteratively refines its analysis over several minutes, saving relevant data and initiating new searches based on what it learns. This process culminates in a detailed research brief with organized summaries and source links.
The reports can be viewed within Gemini apps or exported to Google Docs for editing. Currently, the feature is only available in English on desktops and mobile web, with plans to expand to Gemini mobile apps early next year.
Still, Deep Research raises some concerns. Critics emphasise risks for education, as reliance on AI for brainstorming and problem-solving could erode students’ critical thinking skills. Some studies already link the heavy use of AI tools like ChatGPT to increased procrastination and lower academic performance.
Furthermore, publishers are facing challenges. By aggregating data from multiple sources, Deep Research could diminish website ad revenue, exacerbating issues caused by Google’s AI-generated Search summaries. Since their introduction, publishers report a 5–10% drop in traffic, with potential revenue losses exceeding $2 billion.
Google asserts that Deep Research can help users discover resources they might not find. However, the balance between improving user access and preserving publisher viability remains uncertain.
Gemini 2.0 Flash
In addition to Deep Research, Google has released Gemini 2.0 Flash, an experimental model designed for faster, more accurate chat interactions. This version improves task performance for all Gemini users but is not yet compatible with all features. A full rollout is expected in January.