OpenAI Responds to DeepSeek Hype with 'Deep Research' in ChatGPT Agent
Artificial intelligence firm OpenAI has recently announced the launch of Deep Research, a powerful AI agent integrated into ChatGPT. This new agent is capable of independently conducting extensive web research, analyzing data, and compiling detailed reports, tasks that would normally take humans hours or even days to complete.
According to OpenAI's announcement on February 2, Deep Research is specifically designed for professionals in fields such as finance, science, policy, and engineering, as well as individuals seeking thorough and reliable research assistance.
Competition in AI Research
The release of Deep Research by OpenAI comes at a time when the AI research landscape is becoming increasingly competitive. Chinese-based DeepSeek has emerged as a significant player, challenging the economic dynamics of AI by claiming to match the intelligence of ChatGPT at a more affordable price point.
OpenAI's CEO, Sam Altman, took to Twitter to express excitement about the new tool, describing it as a "superpower" that provides access to experts on demand. Altman also shared a personal anecdote where Deep Research helped him locate a rare NSX car in Japan after struggling to find it manually.
Accessing Deep Research
Users can access Deep Research through ChatGPT by selecting it in the message composer, entering a query, and providing any relevant files or spreadsheets for additional context. The AI then leverages its capabilities to analyze numerous sources within a timeframe of five to 30 minutes, offering a real-time progress tracker before generating a comprehensive report with proper citations.
Challenges and Future Developments
Despite its advanced functionality, Deep Research is not immune to limitations. OpenAI acknowledges that the model may occasionally generate inaccurate information or struggle to differentiate between reliable sources and misinformation.
Currently, the service is available to Pro users with a limit of 100 queries per month, with plans to extend access to Plus and Team users soon. OpenAI is also working on an enhanced and more cost-effective version to cater to the needs of paid users.
To showcase Deep Research's capabilities, OpenAI highlighted its performance on Humanity's Last Exam, an AI benchmark assessing expert-level knowledge across various fields. The o3 model of Deep Research achieved an impressive 26.6% accuracy, surpassing competitors such as DeepSeek-R-1 and OpenAI's GPT-4o.
Future of AI-driven Knowledge Discovery
The introduction of Deep Research by OpenAI coincides with Google's rollout of a similar feature called "Deep Research" for its Gemini AI model, targeting Gemini Advanced users. This move underscores the intense competition between AI industry leaders in shaping the future of autonomous knowledge discovery, signaling a significant shift from the chatbot era that marked the inception of AI technology.
