Meta has acknowledged that artificial intelligence is playing an increasingly important role in keeping users on Instagram for longer periods by delivering smarter and more personalized content recommendations.
During its second-quarter earnings announcement, the company said recent improvements to its AI-powered recommendation systems have led to double-digit growth in the amount of time users spend on Instagram compared with a year ago.
Smarter Content Recommendations
Instagram’s recommendation engine no longer relies primarily on engagement signals such as likes or watch time.
Instead, Meta now uses large language models (LLMs) to analyze the actual content of public posts and Reels, understand their topics and context, and predict which content individual users are most likely to enjoy.
According to the company, every public Reel and post on Instagram is now analyzed by AI before being recommended to users.
Longer Sessions, Higher Engagement
Meta said the latest improvements to its Reels ranking system have resulted in longer user sessions, increased content sharing, and higher engagement across the platform.
The company is now expanding the same AI-driven recommendation approach to Instagram’s main Feed and additional areas of Facebook.
Meta is also deploying a new family of AI models, known as Muse, to analyze videos, classify their topics, and automatically generate summaries.
The Trade-Off of Personalized AI
Meta argues that smarter recommendations benefit both creators and advertisers by helping content reach more relevant audiences and increasing engagement.
For users, however, the company’s disclosure has renewed discussions about the impact of AI on digital habits.
As recommendation systems become increasingly accurate at predicting what people want to watch next, many users may find it more difficult to stop scrolling.
The announcement highlights how artificial intelligence is evolving from a tool for improving user experience into a core technology for maximizing engagement across Meta’s platforms.
