OpenAI unveils new feature for ChatGPT

The company OpenAI has unveiled a new feature for ChatGPT called Deep Research. This tool allows the artificial intelligence to operate autonomously, planning and executing complex, multi-stage research. The Verge reports on the innovation.

Deep Research offers enhanced capabilities for searching, data analysis, and real-time adjustment of search strategies. Unlike standard text generators, this feature displays the working process in a side panel, where users can view sources and analysis results. Context can be provided by attaching questions in text form, as well as images, PDF documents, and tables. Responses to queries are provided within 5–30 minutes. In the future, OpenAI plans to integrate charts and images into research results.

However, despite its powerful functional capabilities, OpenAI warns of certain limitations of Deep Research. These include the possibility of "hallucinations" (invented facts), mixing verified information with unconfirmed rumours, and errors in assessing confidence in conclusions.

The development of agents like Deep Research is becoming a key direction for OpenAI and other technology companies. The tool is oriented towards analytical work and demonstrates its ability to respond to queries using tables and lists. In one example, the AI provided an answer to a query about changes in retail over the past three years.

This tool became available shortly after the launch of OpenAI Operator, which uses a web browser to perform various tasks. Google, for its part, is testing a similar project called Project Mariner, although it is not yet available to the general public. Deep Research is already available in the Pro subscription version.

ChatGPT Pro users can make up to 100 queries per month, and OpenAI will also offer limited access to Plus, Team, and Enterprise subscribers. Since Deep Research requires significant computational resources, the company plans to expand limits for all paid subscribers in the future.