Deep Research: OpenAI’s AI-Driven Feature for Enhanced Information Retrieval

Deep research has been announced by OpenAI as a revolutionary AI-powered tool aimed at improving information retrieval and citation accuracy in ChatGPT. Once available only in a limited preview, this feature is now accessible to all users of ChatGPT Plus, Team, Edu, and Enterprise, vastly increasing its availability.

The update was conveyed through a series of posts on X (formerly Twitter), showcasing significant enhancements to the tool. OpenAI’s CEO Sam Altman expressed that it is “one of the favourite things they have ever shipped.”

Improvements to Deep Research

Since its initial release, deep research has undergone various enhancements for greater accuracy and user friendliness. Notable updates comprise:

• Incorporation of embedded images with proper citations in the responses for a more informative visual experience.
• Enhanced capability to understand and reference uploaded files, leading to improved context retention.

Usage Limitations on Deep Research Queries

In order to manage demand across various user tiers, OpenAI has established monthly limits on deep research queries:

• Users of ChatGPT Plus, Team, Enterprise, and Edu are eligible for 10 deep research queries each month.
• Pro users enjoy access to a more generous limit of 120 deep research queries per month.

System Card Release and Expert Training

Moreover, OpenAI has unveiled a system card that details the construction of deep research, its functionalities, potential risks, and implemented safety measures.

OpenAI stressed that deep research was developed with insights from numerous domain experts, honing its ability to provide precise, well-cited, and trustworthy responses. This initiative aligns with OpenAI’s mission to ensure that AI-assisted research remains reliable and enriching.