Photo: Tada Images/Shutterstock.com
OpenAI is planning to introduce a new memory feature for its ChatGPT platform that enables the AI to remember and recall information from past conversations. The company is currently testing with a limited group of free and Plus users, with plans for a wider rollout soon.
Users can instruct ChatGPT to remember specific details or allow it to autonomously pick up information over time, resulting in an AI that becomes more contextually aware and responsive with each interaction.
As per the company, this feature will bring about significant improvements in personalisation and efficiency within the platform.
A key aspect of this memory feature is the user’s control over its utilisation. OpenAI emphasises that users are in complete control over this feature and can turn it off any time they want. Users can also instruct ChatGPT to forget specific pieces of information or manage memories through settings, offering a level of control and flexibility to address the obvious security concerns.
“If you want ChatGPT to forget something, just tell it. You can also view and delete specific memories or clear all memories in settings (Settings > Personalization > Manage Memory). ChatGPT’s memories evolve with your interactions and aren’t linked to specific conversations. Deleting a chat doesn’t erase its memories; you must delete the memory itself, ” said OpenAI.
The company has also taken measures to assess and mitigate biases, ensuring ChatGPT refrains from proactively remembering sensitive information unless explicitly requested by the user.
Enterprise and Team can also use the memory feature as it allows GPT to learn and adapt to individual preferences and styles. For instance, the AI can remember tone, voice, and format preferences for tasks such as blog post drafting, coding and monthly business reviews.
OpenAI has also revealed plans to extend the memory feature to custom GPTs. GPTs, with their memories, can offer users a more personalised interaction. Builders will have the option to enable memory for their GPTs, creating a tailored experience that remembers user preferences across different AI models.
In the News: Revenge RAT spread camouflaged by legitimate tools
