OpenAI has introduced a new feature called "custom instructions," enabling ChatGPT users to tailor their AI chatbot responses to their specific needs and preferences.
Although currently limited to Plus subscribers, OpenAI CEO Sam Altman mentioned that it will be available more widely in the near future.
With "custom instructions," users can provide additional information, such as their location, occupation, hobbies, and goals, to guide ChatGPT's responses.
The feature allows users to set parameters, including the level of formality in the responses, how the chatbot addresses the user, and the degree of opinion expressed in the answers.
Early testers of the feature have already utilised it creatively.
Some integrated country-specific legal references, maintained a consistent coding format, adjusted ChatGPT's responses based on user-defined expertise levels, and used character sheets to ensure consistent portrayals of ongoing interactions.
OpenAI reassures users that ChatGPT will consider these custom instructions for all future conversations.
Once provided, the model will remember the preferences, eliminating the need to restate them in each interaction.
In April, OpenAI addressed privacy concerns by allowing users to opt out of data usage for training ChatGPT.
Should users choose to opt out, the custom instruction feature will be disabled accordingly.