AI Lets Users Decide What Appears in Their Feeds

Facebook is introducing a major change to how users control the content they see. With new AI-powered customization tools, people can now communicate directly with Facebook’s recommendation algorithms using natural language, telling the platform exactly what they want to see more or less of.

At the center of this update is “Your algorithm,” a new feature that allows users to add and manage their interests in Facebook Feed and Reels.

Instead of relying exclusively on behavioral signals such as likes, comments, and viewing habits, Facebook is incorporating large language models (LLMs) to interpret users’ expressed preferences, context, and intentions.

The most significant development? Recommendations can now adapt in real time, even within the same session.

Facebook is also updating its three-dot menu and introducing more conversational feedback options, making it easier for users to influence what appears in their feeds.

The announcement signals an evolution in Facebook’s recommendation system: from algorithms that primarily infer user interests through behavior to systems that can also understand what users explicitly ask to see.

Key Points

1. Facebook introduces “Your algorithm” for direct content customization.

Users can now access a dedicated space to add and manage the topics they want to see in Feed and Reels. The feature is being rolled out through the three-dot menu.

2. Natural language becomes a new way to influence recommendations.

Instead of relying only on predetermined feedback options, users can describe their interests in their own words. Facebook’s AI interprets these instructions and adjusts recommendations accordingly.

3. Recommendations adapt in real time.

One of the most notable advancements is the ability to update recommendations during the same browsing session. Users no longer have to rely exclusively on their past engagement behavior to influence future content.

4. “Interested” and “Not interested” feedback becomes more conversational.

Facebook is expanding its existing feedback system. After users select either option, they may receive follow-up questions asking them to explain what they liked or disliked.

5. Large language models are changing how content ranking works.

Facebook is integrating LLMs into its recommendation systems to better understand the meaning, context, and intent behind users’ preferences.

6. Behavioral signals remain important.

Likes, comments, watched content, and thousands of other signals will continue influencing recommendations. The new system combines these existing signals with explicit natural-language feedback.

7. Customization tools become easier to access.

Facebook has updated the three-dot menu on posts to make its new personalization features more accessible.

Key Quotes

Here are five important statements directly from Facebook’s announcement.

“Now we’re putting you in the driver’s seat.”

This captures Facebook’s central message: giving users more direct control over their content recommendations.

“New AI-powered tools let you tell Facebook, in your own words, what you want to see in Feed and Reels, and your recommendations respond right away.”

This highlights the shift toward conversational customization and immediate responses.

“Now, large language models are helping us take a genuine leap forward.”

Facebook identifies LLMs as the technology behind this evolution in recommendation systems.

“LLMs can incorporate the interests, context, and intent behind natural language, unlocking an algorithm you can actually talk to.”

This explains why natural-language input represents a significant advancement over traditional feedback mechanisms.

“Tell it what you want more or less of, and your recommendations respond in real time, adapting within the same session to reflect what matters to you right now.”

This is particularly important because it establishes that users’ expressed preferences can influence recommendations immediately, rather than only gradually over time.

Implications

1. Facebook is moving toward a more conversational recommendation experience.

The introduction of natural-language interactions changes the relationship between users and recommendation algorithms. Rather than expecting users to communicate their interests primarily through engagement, Facebook now allows them to express preferences directly.

2. Content discovery could become more closely aligned with immediate user interests.

Because recommendations can adapt within the same session, what users see may become more responsive to their current interests rather than being shaped predominantly by historical behavior.

3. User intent becomes an additional input for content ranking.

The integration of LLMs allows Facebook to interpret the context and meaning behind written feedback. This adds a new dimension to its existing ranking systems, which already analyze thousands of behavioral signals.

4. Feedback may become more detailed and actionable.

Follow-up questions associated with “Interested” and “Not interested” give users opportunities to explain their preferences more precisely. This could help Facebook distinguish between different reasons someone enjoys or dislikes a particular post.

5. Content recommendations will continue combining explicit preferences with observed behavior.

The announcement does not suggest that Facebook is abandoning engagement-based ranking. Instead, natural-language feedback will work alongside existing behavioral signals to refine what users see.

6. The change introduces a new consideration for content creators and marketers.

As Facebook incorporates explicitly stated user interests into its ranking systems, content relevance may increasingly involve how well recommendations match what people actively say they want to see.

The announcement does not explain how these changes will affect organic reach, engagement rates, or content distribution for businesses. Those outcomes remain unclear.

The bigger picture: Facebook is introducing an algorithm that does more than observe what people do. It can also interpret what they say they want. The result is a recommendation experience designed to become more responsive, personalized, and interactive.

Source: https://about.fb.com/news/2026/10/introducing-your-algorithm-facebook/