OpenAI has introduced a groundbreaking feature to its ChatGPT platform, revolutionizing the capabilities of AI chatbots. Unlike previous versions that lacked continuous memory and reset after each conversation, ChatGPT now has the ability to remember users from one conversation to another. Known as “custom instructions,” this opt-in beta feature enables users to set unique parameters that persist between chats.
Custom instructions open up a world of possibilities for personalized interactions with the AI chatbot. For instance, users can specify their profession, such as being a third-grade teacher, and the system will tailor its responses to suit the needs of students. Similarly, users can provide information about the size of their family, enabling the chatbot to provide accurate ingredient lists for recipes.
While this feature is currently available to ChatGPT Plus subscribers as a beta offering, it is expected to be rolled out to third-party apps that utilize ChatGPT as their underlying code. This is particularly advantageous for mobile phone users who often find repetitive instructions cumbersome when using virtual assistants via voice commands. However, OpenAI emphasizes that this feature aims to streamline queries rather than replicate the all-encompassing anticipatory personal AI assistant portrayed in the movie “Her.”
Privacy concerns are a legitimate consideration with this new feature, leading OpenAI to undertake rigorous testing during the beta phase. By subjecting the system to real-world usage scenarios, the company aims to identify and address any potential issues. Moreover, incorporating additional layers of instructions adds complexity to the queries, increasing the risk of the AI generating inaccurate or fabricated responses. Users must bear in mind that this is a pre-release beta version and should not expect flawless performance.
The custom instructions settings tab adheres to the same ethical guidelines as the AI chatbot itself, preventing the system from engaging in inappropriate or harmful behavior. OpenAI provides an example of attempting to include a custom instruction requesting tips on murdering people, which would be rejected by the system. Furthermore, the platform ensures the removal of personal information that has the potential to identify users. This safeguarding of user privacy is commendable, but it also limits the extent to which tech companies can harness valuable data for the creation of fully immersive digital assistants.
OpenAI’s release of this update is set to occur today, but it is currently exclusive to paying ChatGPT subscribers. Unfortunately, users in the UK and EU are temporarily unable to access this feature, though OpenAI plans to expand availability to these regions in the near future.
In conclusion, OpenAI’s ChatGPT platform has taken a major leap forward with the introduction of custom instructions. This innovative feature enables the chatbot to retain memory between conversations, providing a more customized and efficient user experience. While privacy concerns and potential complexities remain during this beta phase, this development sets the stage for a future where AI-based personal assistants can better cater to individual needs and preferences. As this technology evolves and becomes more accessible, it is crucial to strike a balance between personalization and safeguarding user privacy.
