OpenAI has officially launched its latest feature, called 'Deep Research,' which empowers the popular ChatGPT platform to perform complex, multi-step online research on behalf of its users. Designed for individuals across various fields such as finance, science, policy, and engineering, this innovative tool is intended to save users hours of manual research by synthesizing information from hundreds of online sources quickly and efficiently.
Launching on February 3, 2025, OpenAI’s 'Deep Research' allows users to input prompts or questions, and the AI generates comprehensive reports akin to those crafted by professional researchers. This step reflects OpenAI's commitment to developing tools aimed at making knowledge work more accessible and efficient for everyone. "Deep research is OpenAI's next agent. You give it a prompt, and ChatGPT will find, analyze, and synthesize hundreds of online sources to create a comprehensive report at the level of a research analyst," OpenAI stated.
Sam Altman, the CEO of OpenAI, emphasized the tool's capabilities by saying, "This is like a superpower. It can go use the internet, do complex research and reasoning, and give you back a report." The efficiency of 'Deep Research' is impressive, with the potential to complete research tasks within 5 to 30 minutes, drastically reducing the time normally required for thorough investigations.
Powered by OpenAI's forthcoming o3 model, 'Deep Research' is optimized for web browsing and data analysis, enabling the AI to interpret vast amounts of text, images, and PDFs from the internet. The model is also adept at adapting its line of inquiry based on the information it encounters, ensuring the research is relevant and comprehensive.
To utilize 'Deep Research', users simply select the feature from ChatGPT's model selection menu and enter the desired query. They also have the option to attach relevant files or spreadsheets, which allows the AI to tailor its research using the provided data. Once the research is underway, users can track the progress through a sidebar, which summaries the steps taken and the sources consulted.
This functionality is particularly beneficial for those needing detailed reports, whether for market research, policy briefs, or even personalized product recommendations. The AI provides clear citations and reasoning summaries to help users verify authenticity and reliability, making 'Deep Research' not only efficient but also authoritative.
What sets this feature apart is its multi-step research capability. It goes beyond simply answering queries and digs deep, functioning as both researcher and analyst. Reports generated are detailed and well-documented—prerequisites for anyone involved in rigorous academic or professional fields.
OpenAI has confirmed the feature's rollout to ChatGPT Pro users initially, with future plans to extend accessibility to Plus and Team users. The outputs generated by 'Deep Research' currently consist of text-based reports, but OpenAI has announced the integration of images, data visualizations, and additional analytical outputs coming soon. This expansion will undoubtedly augment the usefulness of the tool.
Despite the powerful enhancements 'Deep Research' provides, OpenAI recognizes its limitations. Early tests indicated potential challenges, including the AI's tendency for 'fact hallucinations'—when it generates incorrect information—and difficulties distinguishing reliable sources. Therefore, fact-checking by users remains imperative, as the AI may struggle with confidence calibration and accurately conveying uncertainty.
One of the major tests for 'Deep Research' came from 'Humanity’s Last Exam,' which included over 3,000 expert-level academic questions. Impressively, the o3 model achieved accuracy rates of 26.6%, outperforming other AI tools such as Google's competing product. This testing showcases the advanced reasoning and synthesis capabilities of the tool.
Though it competes with similar offerings from companies like Google, OpenAI's ambitious 'Deep Research' initiative aims to provide thorough, well-researched outputs equipped with citations. Addressing the contemporary challenge of information overload, this feature is positioned to transform how users approach complex research tasks.
Moving forward, OpenAI plans to introduce mobile and desktop versions of 'Deep Research' within the month. The tool aims to support more specialized data sources, including subscription-based platforms, which would broaden its applicability. OpenAI envisions combining 'Deep Research' with other AI tools such as Operator for real-world task execution, highlighting its forward-thinking approach to AI development.
With this launch, OpenAI reaffirms its commitment to pushing the envelope of what's possible with AI-assisted research. The potential for 'Deep Research' to revolutionize how users conduct inquiries is palpable, and users can look forward to enhanced capabilities as the company continues refining the model. Whether boosting productivity for professionals or assisting consumers with informed purchasing decisions, 'Deep Research' stands to make a significant mark on the research-centric tasks of tomorrow.