Meta has unveiled two new enhancements to its AI chatbot aimed at delivering a more tailored user experience.
These updates include a memory function and personalized recommendations based on users’ social media engagement. The features are currently being rolled out on Facebook, Messenger, and Instagram in the U.S. and Canada.
Memory Feature
Meta AI’s memory function allows the chatbot to retain and recall specific user information shared during conversations.
Users can either explicitly instruct the chatbot to remember certain details or rely on its automatic capability to store relevant data.
For example, if a user mentions being vegetarian, the chatbot will remember this and provide only vegetarian meal suggestions in future interactions.
Meta emphasizes that users will receive notifications whenever new information is stored and will have the ability to delete specific memories manually.
The company has not clarified whether sensitive data, such as financial or medical information, can be stored.
Recommendations Based on Social Media Activities
In addition to memory retention, Meta AI now generates tailored suggestions by analyzing data from user profiles and in-app activities across Facebook and Instagram, including likes, comments, and Reel views.
For instance, if a user inquires about family activities for the weekend, the chatbot may recommend attending a concert or dining options based on their home location, browsing patterns, and previously stored preferences.
Managing Meta AI’s Memory
Users have full control over the stored memory. Whenever the AI saves a new detail, a “Memory Updated” notification is issued. To manage or delete this memory, users can:
- Tap “View All” next to the memory notification.
- Review the list of stored details.
- Select and delete specific entries or clear the entire memory.
Importantly, Meta clarifies that memory updates are limited to private conversations and do not apply to group chats.
How Personalized Suggestions Work
Meta AI’s recommendations are derived from various data points, including:
- User Profile Information: Such as home location and demographic data.
- In-App Activity: Engagement with posts, comments, and Reels viewed.
For example, based on a user’s recent Reel views of live performances and family information, Meta AI might suggest concert attendance or brunch reservations.
Privacy Concerns Still Unresolved
Despite the potential convenience of these features, concerns persist about privacy and user consent. Meta has not disclosed whether users can fully opt out of the memory feature or prevent their data from being analyzed for recommendations.
Gradual Rollout
The features are being tested and deployed across Facebook, Messenger, Instagram, and WhatsApp on iOS and Android devices in the U.S. and Canada. Meta has not yet announced plans for expansion to other regions.
As Meta pushes forward with AI-driven personalization, questions surrounding data privacy and transparency are likely to spark further industry discussions.
