For years, the relationship between humans and AI has been intimate and solitary — a single mind consulting a silent oracle. OpenAI is now testing a different premise: that intelligence, artificial or otherwise, may be most useful not in private counsel but in shared company. By opening ChatGPT to group conversations of up to twenty people, the company is quietly asking whether the future of AI assistance belongs not to the individual, but to the collective.
OpenAI Launches ChatGPT Group Chats, Enabling Collaborative AI Conversations
From my assistant to our assistant
Mark
So this is ChatGPT, but for groups. What's the actual difference from just having multiple people in a Zoom call with ChatGPT open?
Mimi
The difference is that ChatGPT is now a real participant in the conversation, not just a tool one person is using. It reads the flow, jumps in when it makes sense, reacts with emoji. It's designed to feel like a group member, not a service.
Luke
But how does it actually decide when to jump in? Is that something OpenAI has explained, or is that still a black box?
Mimi
The reporting doesn't specify the exact mechanism. We know it's powered by the GPT-5.1 Auto model, which adapts to the group's composition, but the decision-making logic for when to contribute isn't detailed.
Mark
What about the privacy piece? If I'm in a group chat with my coworkers, can OpenAI see what we're discussing?
Mimi
OpenAI says group chats don't create memory the way personal chats do, and personal chat memory doesn't bleed into group conversations. But that's about retention, not visibility during the conversation itself.
Luke
Right—and that's an important distinction. The company is saying they're not storing group chat data the same way, but the source doesn't clarify whether OpenAI can access the content while it's happening, or what happens to it after the conversation ends.
Mark
Who's this for, really? Teams? Families?
Mimi
Both. The company describes it as a space for friends, family, or coworkers to plan, make decisions, or work through ideas. It's deliberately broad.
Luke
And it's only in four countries right now—Japan, New Zealand, South Korea, Taiwan. That's a pretty limited test. We don't know yet if this resonates with users or if there are friction points OpenAI hasn't anticipated.
Mark
What happens if someone in the group is under eighteen?
Mimi
Content filters turn on automatically for everyone in that group. Parents can also disable group chats entirely for their kids.
Luke
That's a safety measure, but it also means a parent's choice affects the entire group's experience. That could create some awkward dynamics.
Il Polso
- OpenAI is piloting group chats in Japan, New Zealand, South Korea, and Taiwan, allowing up to 20 users to share a single AI-powered conversation space — a fundamental departure from ChatGPT's one-on-one design.
- The feature introduces a social layer that makes ChatGPT behave less like a tool and more like a participant — reacting with emoji, reading conversational flow, and responding when tagged by name.
- Collaborative tools including web search, image generation, file uploads, and voice dictation are all available within group spaces, raising the stakes for how teams might use AI in real workflows.
- Privacy safeguards are built in: group chats carry no memory, content filters activate automatically when minors are present, and parental controls can disable the feature entirely for children.
- OpenAI is watching early adoption closely before a global rollout, signaling that the company sees this not as a finished product but as a live experiment in how people share AI.
For years, the relationship between humans and AI has been intimate and solitary — a single mind consulting a silent oracle. OpenAI is now testing a different premise: that intelligence, artificial or otherwise, may be most useful not in private counsel but in shared company. By opening ChatGPT to group conversations of up to twenty people, the company is quietly asking whether the future of AI assistance belongs not to the individual, but to the collective.
OpenAI has begun testing a feature that reframes the fundamental nature of ChatGPT: instead of a private exchange between one person and an AI, group conversations allow up to twenty participants to share a single space — brainstorming, planning, and problem-solving with the AI as an active member of the room. The pilot is currently running in Japan, New Zealand, South Korea, and Taiwan, open to both free and paid users, with broader expansion contingent on early feedback.
The setup is deliberately simple. A user creates a group, invites others, and each participant builds a basic profile so the space feels inhabited rather than anonymous. Groups appear in the sidebar as their own dedicated section. If a group is spun off from an existing one-on-one chat, ChatGPT duplicates the thread rather than disrupting the original. The creator holds admin controls; other members can adjust smaller things like the group name or their notification settings.
What distinguishes this from a standard ChatGPT session is the social texture. The AI reads the rhythm of conversation and decides when to speak and when to hold back. It can react with emoji, and users can tag it directly — gestures that make it feel less like a service and more like a presence. Alongside this, the group unlocks collaborative tools: web search, image generation, file and image uploads, and voice dictation. A behind-the-scenes model called GPT-5.1 Auto adapts responses based on each participant's subscription tier, so free and paid users can coexist without friction.
Privacy is carefully bounded. ChatGPT's memory feature does not apply in group settings — nothing learned in a group chat carries forward. If any participant is under eighteen, content filters engage automatically for the entire group, and parents can turn off group chats entirely through parental controls.
The deeper significance is directional. ChatGPT has long been understood as a personal tool — something you use alone. But as AI embeds itself into everyday life, the scenarios multiply: work teams drafting together, families coordinating plans, friends riffing on ideas. OpenAI's group chat feature repositions the AI from "my assistant" to "our assistant," a subtle but consequential shift in how people might come to think about shared intelligence.
OpenAI has begun testing a feature that fundamentally changes how people interact with ChatGPT: group conversations. Instead of a one-on-one chat between a user and an AI, the new group chat function allows up to twenty people to share a single space where they can brainstorm, plan, and work through problems together with the AI as a participant. The company is currently running the pilot in Japan, New Zealand, South Korea, and Taiwan, available to both free and paid users, with plans to expand to additional regions based on early feedback.
The mechanics are straightforward. A user creates a group and invites others to join—no subscription tier required. Once inside, each participant sets up a profile with a name, username, and picture so others know who they are. The group appears in the sidebar under a dedicated section. If someone starts a group chat from an existing one-on-one conversation, ChatGPT duplicates that thread, leaving the original untouched. The person who created the group holds admin powers: they can add or remove members. Other participants have lighter controls—they can rename the group or silence notifications.
What makes this feel different from ChatGPT's traditional interface is the social layer. The AI now behaves more like a member of a group chat than a service waiting for commands. It reads the flow of conversation and decides when to contribute or stay quiet. It can react with emoji, just as a person would. Users can tag it directly, the way you'd mention someone by name in a group text. These touches blur the line between ChatGPT and a social messaging app.
Beyond the social elements, the group chat unlocks a suite of collaborative tools. Members can search the web together, upload images and files, generate images, and use voice dictation. Behind the scenes, OpenAI is using what it calls a "GPT-5.1 Auto" model that adapts to the group's composition. If one person holds a paid subscription and another uses the free tier, the system can serve different responses to each, optimized for their plan's capabilities.
Privacy and safety are woven into the design. ChatGPT's memory feature—which normally learns from a user's personal conversations to provide better context over time—does not carry over into group chats. No memory is created from group conversations either. If anyone in the group is under eighteen, content filters activate automatically for everyone in that space. Parents and guardians can disable group chats entirely for their children through parental controls.
The shift is significant because it signals where conversational AI is heading. For years, people have thought of ChatGPT as a personal assistant—a tool you consult alone. But as AI becomes woven into daily routines, the use cases are multiplying. Work teams need to collaborate on presentations. Families plan vacations together. Friends brainstorm ideas. In each scenario, having a shared AI participant could become natural. OpenAI's move transforms the relationship from "my assistant" to "our assistant," a change that could reshape how groups approach decision-making and creative work. The company is watching how users respond in these four pilot regions before deciding how quickly to push the feature worldwide.
Citazioni salienti
Bringing your friends, family, or coworkers into a shared space to plan, make decisions, or work through ideas together— OpenAI blog post describing the group chat functionality
Vuoi la storia completa?
Leggi l'originale su India.com ↗
Technology