ChatGPT Tries Going Social. The Group Chat Era Has Arrived.
Chat apps have spent the last three years borrowing ideas from AI. Now the traffic is flowing the other way.
OpenAI has begun rolling out group chats to every tier of ChatGPT, a move that shifts its signature product from a private thinking partner into something closer to a shared workspace for planning, debating, and co-creating.
The feature, released globally after a short pilot in Japan and New Zealand, is designed to let up to twenty people and one AI work together inside a single thread.
OpenAI frames it as a natural evolution. ChatGPT can now sit quietly in the corner until tagged, step in with a summary when needed, or react with an emoji like everyone else. It is meant to feel more like a participant than a tool.
The premise is straightforward. OpenAI wants ChatGPT to help groups coordinate trips, plan events, tackle research problems, or write documents together.
Each person keeps their own memory and settings, and invites are required before anyone joins. Adding someone to an existing chat creates a fresh conversation rather than rewriting the past. Profiles are minimal, with a name, username, and photo.
OpenAI says these threads aim to make collaboration less dependent on email chains or fragmented messaging apps. Tap the people icon, add participants or share a link, and the group can get started. Under the hood, the model deciphers when to volunteer information and when to hold back.
The timing is not accidental. The launch follows the debut of GPT 5.1, which introduced Instant and Thinking modes, and comes on the heels of Sora, a social video experiment built around generative clips of users and their friends. Put together, they sketch a company testing what AI looks like when it becomes part of interpersonal dynamics rather than a private assistant.
OpenAI hints that this is an early step in a longer shift. In a note to TechCrunch, it said it expects ChatGPT to become an active presence in group discussions, helping people plan and take action together rather than acting as a detached reference bot.
If the move succeeds, AI could become a fixture of group decision-making in the same way search engines quietly underpin almost every conversation. If not, it may simply prove that people tolerate AI advice in private but prefer human chaos in public. Either way, the experiment has begun.
