ChatGPT just became not only a solo assistant, but an actual chat app. OpenAI is launching group chats natively, for all including free users, transforming the AI into a true shared space where friends, classmates, and teams can collaborate alongside the model in real time.
Here, a recognizable messaging dynamic — threads, invites, and more than one human participant — is incorporated into the ChatGPT interface. Instead of one-on-one nudging, the AI is now part of a conversation with up to 20 members at once, addressing the group and staying aware of context as the exchange develops.
It’s a small shift with large implications: No longer does ChatGPT merely answer your questions; now, the model can help groups align on decisions, fact-check claims in real time, and document outcomes as you proceed.
What Changes Inside ChatGPT Group Conversations
Group chats exist within the same interface you’re already accustomed to. Tap the group icon in the top right of a new chat or an existing chat to create a room, generate an invite link, and share it with people. Before joining, participants fill out a slim profile that allows the AI to speak to them by name and achieve some continuity in long threads.
There is, for now, a hard limit of 20 people in a conversation. All messages — including yours, your teammates’, and those sent by the AI — go into a collective timeline so the model has as good an understanding about what you know as your group does. These rooms are safeguarded in a unique section of the sidebar so even if you have a private chat, the room will be placed with other public channels.
It is available to premium and free tier users worldwide as long as a user logs in. That’s important because it tears down a paywall to the core experience, enabling everyone from study groups to community organizers and small teams of hobbyists working on open-source projects to get AI assistance without having to pay for a subscription.
How Group Chats Work in the New ChatGPT Experience
Think of the AI as a collective partner.
In a product sprint, one person can paste in user feedback, another can add some kind of draft spec, and the AI will summarize themes that are emerging from that, propose potential requirements, and flag open questions or call for missing information.
When friends are getting ready to take a trip, the model can also help them run through possible itineraries by comparing routes and pointing out conflicts.
Since the AI has access to the entire thread, it’s able to map divergent points of view, cite prior messages, and create artifacts that are group-ready — bulleted summaries, decision logs, or code snippets — without each participant having to recapitulate context. Latecomers can scroll up to read the rationale and ask the AI for a two-minute fill-in of what they missed.
It also reduces “prompt silos.” Instead of a dozen chat threads at various levels of completion, one single chat room can become the thread itself — a canvas collecting sketches — a conversation that becomes the day-to-day documentation of the evolving project.
Why This Matters for Messaging, Teams, and Collaboration
Group chats push ChatGPT into the realm of everyday messaging and collaboration apps. It’s not going to immediately replace Slack, WhatsApp, or Teams, but it does collapse two workflows — talking with people and forcing an AI to act — onto a single screen. The development reflects how workplace tools are integrating generative AI into shared spaces rather than sequestering it off as a standalone bot.
It seems there’s an audience itching to test it out. OpenAI has said ChatGPT attends to more than 100 million weekly active users, meaning any new feature gets immediate scale. The Pew Research Center has shown that an increasing percentage of U.S. adults have tried ChatGPT at least once, and its use is expanding most quickly among students and knowledge workers — the populations who have the most to gain from collaborative prompting.
And the design is also consonant with actual behavior. People already cut and paste AI outputs into group threads to argue, refine, or fact-check. When you bring the model into the room, it reduces copy-paste friction and provides everyone with a window into the exact same instructions/sources.
Privacy and Limits for Shared ChatGPT Group Rooms
And since it’s a shared space, anything you type will be visible to everyone in that room and continue to live there in the room history. OpenAI’s standard controls apply: consumer accounts can control whether chats are saved to history and used for training models, while enterprise offerings generally opt out of training by default, the company’s documentation says.
There are limits in place to manage expectations, after all. With a limit of 20 people, rooms can’t devolve into chaos, and the AI still has to obey platform safety and accuracy regulations. For sensitive work, teams must validate outputs and ideally use accounts with compliant data governance.
What Comes Next for ChatGPT Group Chat Features
Expect rapid iteration. Naturally, the next steps would be more fine-grained control from there — pinning shared agendas, doling out tasks, or exporting summaries in some way — as well as better integration with voice and real-time search so that the AI could schedule meetings and get live data without leaving chat.
For consumers, the headline is simple: ChatGPT can now chat for real. If you’ve been using it as only a research assistant, bringing people into the room shifts the dynamic — and, usually but not always, how much progress can be made with AI.