How AI 'Makes' Users Spend Time on Instagram

Meta CEO Mark Zuckerberg said that in the second quarter, time spent by users on Instagram grew by double digits thanks to improved recommendation algorithms in the feed and Reels. The number of sessions increased by 15 basis points, with reposts and in-app time seeing particular growth. The recommendation system uses a large language model that analyzes posts by topic and tone. The company plans to expand these technologies to Facebook.

Meta CEO Mark Zuckerberg explained why Instagram users are spending noticeably more time in the app. In the second quarter, audience time in the service grew by double digits. According to him, this was driven by improved recommendation algorithms in the main feed and the Reels section. Zuckerberg said that recommendations are becoming more personalized: the platform shows more fresh content and gives users more control over what they see. Instagram Reels introduced a new architecture that takes into account a more detailed history of user interaction with the platform, allowing for more precise content selection. Thanks to these changes, the number of sessions on Instagram increased by 15 basis points. Particularly notable growth was recorded in the number of reposts and the time users spend in the app. The recommendation system uses a large language model that analyzes all public posts and Reels videos by topic and tone. Meta intends to expand the use of such technologies to Facebook as well. Earlier, Cursor reported that Meta introduced new AI-based features in Instagram, expanding image processing capabilities.