A user asks OpenAI to turn ChatGPT into a way for people to meet each other

A proposal from the community forum

A post on OpenAI's community forum, as community.openai.com showed, asks the company to consider a new direction for ChatGPT. The author, a user from Sudan, suggests adding a feature that lets people around the world get in touch with each other, with the AI acting as a safe go-between.

The stated purposes are broad: getting to know people, exchanging information, and connecting for different reasons. The author's own title for the idea translates roughly as AI serving as a bridge between people.

This is a suggestion from one person, written as a message to the OpenAI team. It is not a product announcement, and it has no company backing. Our newsletter flagged it as a community idea for that reason.

How the feature would work

The post lays out a simple flow in five steps.

• An optional directory lists users who are willing to be contacted.
• A user who wants to reach someone asks to do so through ChatGPT.
• The system sends the other person a notice that someone wants to get in touch.
• That person is free to accept or refuse.
• If they accept, the system lets the two begin communicating.

The key design choice is consent on both sides. Nobody would be listed unless they chose to be, and nobody could be contacted without being able to say no. That is closer to a request-and-approve contact list than to an open feed where strangers can see everything.

The author lists the hoped-for results: easier contact between ChatGPT users in different countries, shared experience and ideas, and a new route to collaboration and knowledge sharing.

Privacy and safety are the hard part

To its credit, the post names the safety questions itself. It says the feature should depend on both parties agreeing. It also calls for a way to verify users, protection of personal data, the ability to turn down requests, and a way to block any user. No one's information, it says, should be revealed without their approval.

Those are sensible starting points, but the post gives no detail on how they would be delivered. Verifying identities at global scale is difficult. So is moderating private conversations between strangers, and so is handling harassment or scams once people can reach one another.

There is another question our newsletter raised: who sees what you have written to an assistant. Many people use ChatGPT for personal or work matters. Any feature that links accounts would need very clear rules so that nothing from those private chats leaks into a social setting.

Why a forum wish is still worth reading

Ideas like this do not signal what a company will build. They do show what ordinary users feel is missing. Here, the author argues that an AI could do more than provide information and answer questions, and could also help people reach each other.

That is a different picture of the product from the one most people have. Today, ChatGPT is largely a one-to-one conversation between a person and an assistant. The proposal imagines it as a hub among many people.

Our newsletter made a practical point as well: if you run a community or a team, the nearer-term version of this demand tends to show up in shared workspaces rather than in a global network. That is our reading, not something the post says.

What to expect next

As of the newsletter's reading, there was no sign of a company reply, and nothing suggests OpenAI is considering the feature. The post simply asks the team to study the idea and weigh it for future updates.

The sensible response is patience. Do not build plans, communities or business models around a feature that has only been suggested. If OpenAI ever does move in this direction, the details to look for are how people opt in, how identities are checked, and what happens to chat data.