OpenAI has enhanced ChatGPT's Deep Research feature with new controls for users to specify which sources the AI can pull information from. The update, rolling out to Pro and Plus users with Free and Go tiers coming soon, adds a redesigned side menu and more connected apps. Users can now direct the AI to only use specific websites like bea.gov or census.gov when compiling reports.

This update aims to give researchers greater control over report accuracy while maintaining the feature's ability to produce "expert-level" analyses. The AI can now handle follow-up instructions mid-task, allowing users to refine their requests in real-time. These comprehensive reports can be viewed in fullscreen within ChatGPT, making it easier to digest the AI's findings.

OpenAI's move comes as Google continues to expand its own AI research capabilities with features like AI Mode and enhanced Gemini functionality. The competition highlights how both companies are racing to perfect AI-powered research tools that can sift through massive amounts of information. While these tools promise speed and comprehensiveness, users should still verify facts independently, as the personal touch remains essential for ensuring accuracy in research.

ChatGPT Go, OpenAI's new $8 subscription tier, leverages the GPT-5.2 model for faster responses and offers 10x more messages than the free version.