OpenAI has launched the long-awaited "conversation branching" feature for ChatGPT users. This new tool allows users to create different scenarios and explore alternative ideas from a single conversation.
Now, users can click on a message during a chat and select "branch in new chat" from the "more options" menu. This copies the current conversation and opens a new chat window, allowing users to work on alternative scenarios without losing the original chat.
Previously, users had to either alter the existing chat or start a completely new one, disrupting workflow and limiting creativity. The new system enables trying out "what if" scenarios without losing context.
Experts say this feature resembles the "Git branch" system used in software development. Just as developers test changes on new branches without disturbing the main code, ChatGPT users can test alternative responses and ideas similarly.
Research shows linear conversations with large language models can cause inefficiencies in tasks like project planning and information gathering. ChatGPT's new feature aims to overcome these limitations.
Anthropic's Claude has had branching for over a year. However, ChatGPT's larger user base means this feature will likely have a greater impact.
Users often see AI as a single "authority," which can limit creativity. Branching emphasizes AI as a tool, allowing testing different viewpoints, disproving or confirming hypotheses.
OpenAI will also launch a parental control feature next month. This follows a lawsuit filed by a family after a young person died by suicide following prolonged chats with ChatGPT.
