Mark Zuckerberg says the company's AI-driven feed and video recommendations have led to an 8% increase in time spent on Facebook and a 6% increase on Instagram
Sarah E. Needleman / Wall Street Journal :
Context & Ripple Effects
Meta has been shifting its feeds from primarily social-graph distribution toward AI-selected content for years. It had projected that recommendations would become a much larger share of Facebook and Instagram feeds, while Reels had already been identified as a major engagement driver.
The reported gains follow Meta's work on a cross-service video recommendation system that produced an 8% to 10% lift in Reels watch time, and extend the earlier claim that AI recommendations helped increase Instagram time spent after Reels launched. The significance is that Meta is now attributing measurable platform-wide engagement gains to that recommendation infrastructure.
First-order effects
- Facebook and Instagram users are being served more AI-ranked feed and video content, with Meta reporting higher time spent on each service.
- Meta gains evidence that recommendation-model investment is improving engagement across its two largest consumer surfaces, rather than only within Reels.
Second-order effects
- The results raise the performance bar for rival social and video platforms: retaining audience attention increasingly depends on ranking systems that can surface relevant content beyond a user's follow graph.
- Creators and publishers become more dependent on recommendation eligibility and audience-response signals, since AI-selected distribution can matter more than an established follower base.
Third-order effects
- If these gains persist, social platforms will compete less on the size of users' social graphs and more on their ability to turn a broad content supply into personalized viewing sessions.
- The shift also concentrates product power in opaque ranking systems, making the trade-offs of engagement optimization—especially where recommendations are adjusted to retain teen users—a more durable governance issue.
The trend: This is one data point in the rise of AI recommendation systems as the core distribution layer for social platforms' feeds and short-form video.