On Thursday, OpenAI appeared to introduce a range of new options to its Custom Instructions feature for ChatGPT. Users flocked to social media to share screenshots of these enhancements, which provided a more tailored approach to the chatbot’s responses, including fields for users to input their nicknames, professions, and personality traits. However, these options seemed to have been removed shortly after their introduction, leading to speculation that OpenAI may have prematurely enabled the feature.
ChatGPT Temporarily Offered Enhanced Customization Features
Several users on X—formerly Twitter—reported the emergence of this new functionality within ChatGPT. This update seemed to expand on the existing Custom Instruction feature, now providing users with additional customization options aimed at refining how the AI generates responses.
Custom instructions serve as foundational tools, reducing the need for repetitive prompts. For instance, if a user prefers responses that are formal and concise while maintaining an analytical tone, they can specify these preferences in the custom instructions menu. This allows ChatGPT to consistently generate responses in line with specified guidelines.
Recent custom instruction options available on ChatGPT
Photo Credit: X/Tibor Blaho
Previously, users could only add instructions as a single text prompt. However, the screenshots indicated that new fields were included, providing users with greater control over how ChatGPT tailors its responses.
The updated feature allowed users to input a nickname and specify their profession. Additionally, personality traits such as “witty,” “opinionated,” or “Gen Z” could also be added to enhance the chat experience.
While it remains uncertain how these traits influence the chatbot’s outputs, attempts by staff at Gadgets 360 to validate the feature were unsuccessful, and it was not appearing on the chatbot at the time this article was prepared.
Moreover, users on X noted that the feature disappeared shortly after its brief introduction. As highlighted by TechCrunch, it seems OpenAI may have unintentionally deployed the feature before it was fully ready for release. The timeline for an official rollout remains unclear.
