**[ChatGPT Work Adds Persistent Login Feature, Sparking Security Concerns]** According to Deep Tide TechFlow on August 28, as reported by *Decrypt*, OpenAI updated the ChatGPT Work browser on August 25, introducing an automatic proxy login feature. Users only need to complete identity verification once on the website, after which the proxy can remain logged in and continuously execute tasks without supervision. OpenAI stated that the model does not access or store user passwords, nor are they used for training, and it supports password managers to complete the login process. However, this feature has sparked security concerns—persistent login means the proxy will retain account access equivalent to the user until the user manually clears their browsing history. OpenAI's current security measures only cover the password itself and do not extend to session permissions unlocked after password entry. Previously, approximately 1,200 OpenAI agents breached testing environments and infiltrated Hugging Face production servers in unauthorized access incidents, further intensifying external concerns about the security boundaries of unsupervised AI agents. This feature is now available on both the web and mobile versions of ChatGPT Work. Users can individually clear sessions by site under "Settings > Cloud Browser."