OpenAI has announced that its new AI tool, Operator, may retain user chats and screenshots for up to 90 days, even if users manually delete them. This retention period is more extended than the 30-day policy for its popular ChatGPT platform.
OpenAI justifies the Operator’s extended retention by citing the need to monitor potential abuse and improve fraud detection, as the technology is still in its early stages. According to the company, this approach is essential for ensuring the tool’s security while maintaining user control over their data.
Operator, unveiled as part of a research preview for ChatGPT Pro subscribers paying $200 per month, is an AI-powered agent with a built-in browser. It is designed to perform various automated tasks, including booking travel accommodations, making restaurant reservations, and shopping online.
Users can select from task categories such as dining, delivery, shopping, and travel. The tool’s built-in browser captures screenshots to help it understand how to interact with websites, including identifying buttons and completing forms.
However, screenshots are not taken in sensitive scenarios, such as when passwords are required, where the system enters a “take over” mode.
Despite these precautions, some users may have concerns about data privacy, particularly given the longer retention period and the potential for authorized OpenAI personnel or trusted service providers to access data for investigations or legal purposes.
OpenAI has emphasized that the policy is crucial for monitoring abuse vectors and improving the safety of the tool in its developmental phase.
Operator’s functionality aims to streamline tasks, offering a glimpse into the future of AI agents that can perform online activities independently. However, the extended data retention period may prompt questions about the balance between technological innovation and user privacy. As OpenAI continues to refine its tools, maintaining trust and transparency with users will likely remain a critical focus.
