Meta AI Uses Past Interactions to Personalize Facebook and Instagram Feeds

Meta AI Uses Past Interactions to Personalize Facebook and Instagram Feeds

Meta AI now provides personalized features that remember your past interactions on Facebook and Instagram to improve your experience.

Meta has shared updates about its AI assistant, Meta AI, which now includes more personalized features and the ability to remember things. By using information from Facebook and Instagram profiles and previous interactions, these improvements aim to make Meta AI more personalized for each user.

The “Memory Boost” feature, added on January 27, 2025, lets the assistant remember what you like and give you personalized suggestions.

For example, if someone is vegan, the assistant can recommend recipes or nearby restaurants. The assistant can now use your account information to suggest events or restaurants based on where you are and what you’ve recently looked at on Instagram.

Meta AI on Llama 3.3

Meta AI is based on Llama 3.3, a large language model from Meta. You can use it on Messenger, WhatsApp, Instagram, and Facebook. It was first introduced in September 2023 at the Meta Connect event.

The assistant can search the internet in real-time using Bing and works with Ray-Ban Meta smart glasses, allowing users to ask questions about what they see through the glasses.

These updates show that Meta wants to make its assistant a top personalized AI tool while also taking user privacy concerns seriously. Users can look at and remove their saved data whenever they want.

This change fits with a larger trend where tech companies are working to make AI assistants that can hold better conversations and understand context.

 

Leave a Reply

Your email address will not be published. Required fields are marked *

OpenAI Unveils ChatGPT Gov for U.S. Agencies, Backed by Microsoft Previous post OpenAI Unveils ChatGPT Gov for U.S. Agencies, Backed by Microsoft
Meta Sets Up Four 'War Rooms' to Evaluate DeepSeek’s AI Model Next post Meta Sets Up Four ‘War Rooms’ to Evaluate DeepSeek’s AI Model