Google recently made a major announcement in the field of AI, launching a new tool called Deep Research that is revolutionizing academic research and education. This tool, which runs on the Gemini bot, is capable of searching through hundreds of websites in just a matter of minutes. According to Google Deepmind CEO Demis Hassabis, this is a significant step towards the next generation of agent-based systems that can act as universal digital assistants.
Zoubin Ghahramani, VP of research at Google Deepmind, shared that it has been his dream to bring these “intelligent agents” to life in order to simplify research. This new tool has been hailed as one of the most impressive and “Google-y” uses of AI to date by experts like The Wharton School’s professor Ethan Mollick, who had early access to the tool. While there are some limitations due to paywalls around academic sources, Mollick noted that the content is accurate at an undergraduate level.
Previously, Mollick also pointed out that other LLMs (large language models) like Anthropic’s Claude Haiku, OpenAI’s o1, DeepSeek’s R1 Lite Preview, and Perplexity also offer reasoning features. However, Google’s Deep Research tool sets itself apart by incorporating search capabilities as well. Perplexity CEO Aravind Srinivas has compared the two tools, stating that while Perplexity Pro is better for regular searches, Gemini’s Deep Research is more suitable for intensive searches.
Tech journalist Dean W Ball shared on X that Gemini’s Deep Research is well-received for policy research, but may face paywalls for medical queries. While other LLMs like Perplexity and ChatGPT have previously challenged Google’s dominance in search, Google’s vast amount of training data and AI talent give it an advantage. However, concerns about hallucinations and inaccuracies in LLMs, as seen with Meta’s Galactica, have raised concerns about the potential for misinformation and its impact on scientific integrity.
In August, Japan-based Sakana AI introduced the “AI Scientist,” a system that uses LLMs to independently conduct research, from generating ideas to writing and reviewing papers, at a cost of under $15 per paper. This showcases how AI is becoming more active and involved in the research process, rather than just being an expert tool.
