According to recent reports, the new social features in ChatGPT are internally codenamed “Calpico” and “Calpico Rooms.” These may allow users to chat one-on-one or in groups, share content, collaborate on ideas or projects, and even integrate AI generation (text, image) within conversations. The Android version is said to be the first testbed for these changes. 
This shift follows OpenAI’s broader trajectory: expanding from pure AI chatbot toward a more integrated “everything app” model. Previously, OpenAI has already experimented with social tools—like the Sora app for AI-video sharing—and internal prototypes of social media platforms. The new messaging features would blur lines between user and creator, between AI and social interfaces. 
Of course, this evolution raises important questions: how will privacy, moderation, and data access be handled? Will conversations be end-to-end encrypted? Will group chats expose users’ private prompts or AI memories? As OpenAI ventures into social territory, balancing utility and trust will be key.