Meta has reported a significant increase in user engagement on Instagram, with individuals now spending more than 10% more time on the platform compared to the previous year. This surge is directly linked to the company's sophisticated artificial intelligence (AI) recommendation algorithms.
How Meta's AI Boosts Instagram Engagement
During its Q2 earnings call, Meta CEO Mark Zuckerberg and CFO Susan Li highlighted the impact of the new AI-powered systems. These algorithms are designed to deeply understand user preferences, leading to more compelling content recommendations across Reels and Feed posts.
Unlike older systems that primarily relied on likes or browsing history, Meta's redesigned recommendation engine utilizes large language models (LLMs). These models analyze all public Reels and Feed posts to discern topics, tone, relevance, and overall quality. This allows the AI to predict which content users are most likely to watch to completion, like, comment on, or share.
Understanding User Behavior for Tailored Feeds
The AI system meticulously tracks various user behaviors within the app, including:
- Videos watched until the end
- Posts liked, shared, or saved
- Reels replayed
- Accounts frequently interacted with
- Longer-term viewing history
By processing these signals, the algorithm creates a highly personalized content stream, making the platform more engaging and, consequently, more time-consuming for users. Meta plans to extend a similar LLM-powered recommendation system to Facebook in the near future.
Scrutiny Amidst Increased Usage
Despite the success in boosting engagement, Meta continues to face scrutiny regarding the addictive nature of its products, particularly concerning young users. The company is currently embroiled in several legal battles addressing the impact of Instagram and Facebook on children and teenagers. Meta has allocated $2.40 billion to cover legal expenses related to these ongoing lawsuits and investigations.