OpenAI launches new ChatGPT agent with ‘deep research skills’

OpenAI has launched a new AI agent to support users conducting thorough, complex research. This advanced tool, called deep research, is built to handle intricate queries and provide detailed, reliable results.

Aimed at finance, science, policy, and engineering professionals, the new AI agent promises to be especially useful for those requiring precise information. OpenAI also sees potential benefits for individuals making careful purchasing decisions, like buying cars, appliances, or furniture.

A powerful research tool

Unlike ChatGPT’s usual quick answers, deep research is tailored for users who need more than a simple summary. OpenAI sees the feature as ideal for those requiring extensive research across multiple websites and sources.

For example, someone researching a specific scientific topic or an engineer looking for precise technical details can now rely on ChatGPT’s deep research capabilities.

The company noted that the feature would be available to ChatGPT Pro users, who will be able to make up to 100 queries per month. More options for Plus and Team users are expected soon, with an expanded rollout for Enterprise users on the horizon.

How OpenAI’s new AI agent works

To access deep research, users select the option in the composer and enter their query. The tool can process inputs such as files and spreadsheets, which are helpful for those seeking answers from a combination of data. This feature is only available as a web experience, though mobile and desktop app integration is expected later in the month. The research process may take five to 30 minutes, with users receiving a notification once the task is completed.

OpenAI is also looking to expand the tool’s capabilities. While outputs are currently limited to text, the company plans to add images, data visualizations, and other analytical elements. Future updates will allow users to access more specialized data sources, including subscription-based services and internal resources.

Ensuring accuracy and limitations

Despite its promising features, there are concerns about the accuracy of the results produced by deep research. While AI is powerful, it is not without flaws. OpenAI acknowledges that the model could still make errors or hallucinations, especially in complex research scenarios. However, the company assures users that every output will be fully documented, with citations and summaries to help verify the information.

To improve reliability, OpenAI is using its recently developed o3 reasoning AI model, which has been trained using reinforcement learning on real-world tasks requiring web browsing and data analysis. This model has shown promising results in testing, outperforming other AI models like Gemini Thinking and Grok-2 in a benchmark evaluation. Nevertheless, OpenAI cautions that the tool may still struggle with distinguishing authoritative sources from rumors and can make mistakes in formatting reports and citations.

While the deep research agent promises to enhance research efficiency, the true test will be whether users will rely on its outputs and perform the necessary checks or use it as a more professional-looking shortcut. Time will tell if this new tool will become a valuable asset for those in need of in-depth, well-sourced information.

Share your love
Facebook
Twitter
LinkedIn
WhatsApp

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

error: Unauthorized Content Copy Is Not Allowed