A ChatGPT user asks for a real home for feature requests

A request, not an announcement

A post on OpenAI's community forum, as community.openai.com showed, is titled "Dedicated Product Feedback & Feature Request System in ChatGPT." It was written by a user. It is not a statement from OpenAI, and nothing in it says the company plans to build what it describes.

The author's complaint is simple. ChatGPT, they say, has no clearly labeled, always-available place to send ideas about the app itself, such as interface changes or new features. They ask OpenAI to create one.

Why the thumbs-down button is the wrong tool

According to the post, people who want to suggest an improvement are often pointed to the thumbs-down button that sits next to an individual answer. That button is meant to say "this reply was not good." The author argues it ends up treating an app suggestion like a complaint about a response, even when the response was accurate and useful.

The post lists several other problems:

• The feedback controls may not appear in every version of ChatGPT.
• Users may have to contact Support just to learn how to submit a suggestion.
• There is no confirmation or reference number.
• There is no way to see whether a suggestion was received, reviewed, accepted, declined or built.

The author gives an example. Someone trying to suggest a navigation improvement through official support was repeatedly told to use a thumbs-down control that did not exist in their browser view. Support might write the idea down, the post says, but that does not necessarily create a formal request that anyone tracks.

What the proposal would build

The post asks for a separate Feedback & Suggestions option reachable from the Help menu, kept apart from the controls for rating answers. It should sit in the same place and behave the same way on the web, Windows, macOS, Android and iOS.

Users would pick a category for each submission: feature request, interface or user-experience improvement, bug report, accessibility improvement, or general feedback. Each category could have a few optional fields, without making the form heavy.

After sending, the user would get confirmation, a unique reference number and a history of past submissions. Each one would carry a status. The core path suggested is Submitted, Under Review, Planned, Implemented. The author adds possible extras: Needs More Information, Declined and Closed.

Two further ideas stand out. Users could opt in to notifications when a suggestion changes status or an issue is fixed. And the tool could show existing suggestions before someone files a new one, letting people back or follow a proposal instead of creating a duplicate.

Privacy and honest promises

The post is careful about privacy. A feedback form should not quietly pull in private conversation content, project names, chat titles or other sensitive details. Screenshots, conversation excerpts, project context and diagnostic information would be added only if the user chooses, with a preview of exactly what will be sent.

It also warns against overpromising. Status changes should reflect real product-team decisions, and the interface should not suggest every submission gets individual review, or will be built, unless those processes actually exist. That matters because a tracker that implies attention it cannot deliver could frustrate users more than having none.

Benefits and a phased rollout

For users, the author expects clearer routes for ideas, confidence that feedback arrived, less confusion between rating an answer and suggesting a change, and less need to go through Support. For OpenAI, the claimed gains are better-organized feedback, easier spotting of repeated problems, fewer duplicate requests and fewer support tickets about how to give feedback. These are the author's expectations, not results.

The post suggests building it in stages. A first version would include the submission screen, categories, privacy controls and confirmation numbers. History, status tracking, duplicate detection and notifications could come later.

The requested outcome is modest: consider building a cross-platform feedback tool with acknowledgment and room to grow into tracking. Whether OpenAI replied is not stated in what we have.