Facebook's AI Engine Supercharges Reels Recommendations

Facebook is deploying a significantly upgraded recommendation engine across its platform, designed to understand user interests faster and deliver more personalized video content. The improvements come as the company reports video watch time has grown more than 20% year-over-year in the US, signaling a strategic push to dominate social video consumption.

The enhanced AI system now surfaces 50% more Reels from creators published that same day, dramatically increasing the freshness of content users encounter. This acceleration in content discovery addresses a common user complaint about stale or repetitive recommendations that has plagued social platforms.

For users who prefer longer-form content, the update brings good news. Reels exceeding one minute now account for a quarter of content from creators with over 10,000 followers and represent more than half of total watch time on Facebook. This suggests the platform is successfully bridging the gap between short-form viral content and more substantial video experiences.

User feedback mechanisms have been streamlined with new "Not Interested" and comment flagging options that directly influence the recommendation algorithm. Each interaction helps fine-tune the system, creating a more tailored experience that Facebook claims has already led to increased watch time globally.

The update introduces "friend bubbles" on Reels and Feed, allowing users to see which content their friends have liked and instantly jump into private chats about shared interests. This social layer aims to recapture the platform's original community-focused roots while leveraging existing social graphs.

AI-powered suggested search now appears on select Reels, enabling users to explore related content without leaving the video player. This seamless discovery feature represents Facebook's broader investment in AI-driven personalization as competition intensifies in the social video space.

Save functionality has also been enhanced, making it easier to collect favorite Reels and posts in one place. Each save action contributes to the algorithm's understanding of user preferences, creating a virtuous cycle of improved recommendations.

These changes arrive as social platforms increasingly compete on recommendation quality rather than just content creation tools. With video consumption patterns shifting toward algorithmically-curated feeds, Facebook's latest moves position it to better compete against TikTok and YouTube Shorts in the battle for user attention.