This update allows Gemini to use information from previous exchanges to provide more accurate answers and support more effective task completion.
This possibility has long been expected in the field of artificial intelligence (AI). Previously, primary language models were like white paper with each new interaction. They have no context other than the current reminders, making it difficult to have complicated or prolonged conversations.
Over time, researchers have developed a method to help AI have short-term memory, allowing it to remember information in a single conversation. This is a big step forward, but there are still many limitations, because when the conversation ends, AI's memory will be erased.
The long-term memory feature that Gemini Advanced has just deployed is considered a breakthrough. Now, Gemini can access and process information from previous interactions, bringing many potential applications, especially for long-term projects such as programming, spelling or research.
Users will not need to re-explain the goals or steps they have taken, as Gemini can refer to previous conversations.
However, AI's ability to store long-term memories also raises concerns about privacy. To address this issue, Google said it has integrated options for users to review, delete, and manage their chat history. Users can also turn the feature off completely if they want.
The new feature is now available in English to Gemini Advanced users through the Google One AI Premium package on Gemini's web and mobile apps.
Google is expected to expand language support and provide this feature to Google Workspace Business and Enterprise customers in the coming weeks.
