OpenAI Implements Custom Instructions Feature in ChatGPT for Select Users
OpenAI has recently introduced a new feature called custom instructions on ChatGPT, their popular AI chatbot. Although currently in beta, this feature aims to enhance user experience by eliminating the need to input the same prompts repeatedly during each interaction with the chatbot.
According to The Verge, custom instructions on ChatGPT serve as a storage system for users’ preferences. OpenAI states that this feature focuses on remembering important details about the users, such as their life stories or other relevant information that the chatbot may require.
This is particularly beneficial for users who often forget the prompts they previously used. With custom instructions, there will no longer be any difficulty in recalling and using their preferred prompts each time they access the chatbot.
For example, if a user enters a prompt related to programming, ChatGPT can save those instructions and automatically suggest them for the user’s subsequent requests. This feature proves especially handy for individuals who write articles and rely on repetitive prompts daily, as ChatGPT can recall the context and provide useful suggestions accordingly.
In response to queries about compatibility with plugins, OpenAI confirmed that custom instructions work seamlessly with such extensions. In the event that a user asks ChatGPT for recommendations on decent restaurants nearby, the chatbot will promptly retrieve the most recent prompt used for that specific purpose.
TechCrunch reports that currently, only Plus subscribers in the United Kingdom and the European Union have access to this beta feature. However, OpenAI may roll it out in other regions in the future.
According to OpenAI, the usage of custom instructions will also aid in improving the model’s performance by teaching it how to adapt responses based on user instructions without going overboard.
If users wish to discontinue using the feature, they can easily opt out by adjusting their data control settings.
OpenAI has imposed a limitation on the length of responses on the platform, capping it at 1,500 characters to ensure optimal performance and efficiency.
In terms of safety, OpenAI has implemented its moderation API to address any potential issues with customized instructions. The company possesses the authority to accept, reject, or enforce instructions based on whether they comply with their existing policies. Consequently, this prevents users from abusing the chatbot or posting hateful content.
Joanne Jang, involved in OpenAI’s model behaviors, revealed that the chatbot does not retain user information once a new conversation is initiated. However, it can recall certain details from previous conversations.
Additionally, Tech Times highlights that ChatGPT and other similar large language models have the potential to flag fake news. However, they are not yet as reliable as human fact-checkers.
In conclusion, OpenAI’s implementation of custom instructions on ChatGPT aims to streamline user interactions and enhance their experience with the chatbot. While limited to select users during the beta phase, this feature has the potential to be accessible globally in the future, offering users a more convenient and personalized AI chatbot experience.
