Meta Adds AI-Powered Facebook Feed Controls With Your Algorithm

Meta has introduced a new AI-powered feature called “Your Algorithm” that allows Facebook users to customise the content recommended to them in Feed and Reels. Announced on October 13, 2026, the tool lets people directly tell the platform which topics they want to see more or less frequently, using their own words. The change gives users another way to influence content discovery instead of relying exclusively on the platform’s interpretation of their likes, comments, viewing activity and other engagement signals.

The feature uses large language models (LLMs) to help Meta’s recommendation systems interpret users’ stated interests and adjust suggestions in real time. Users can access the controls through the three-dot menu on posts and provide feedback through options such as “Interested” and “Not interested.” The launch reflects Meta’s wider effort to make its social platforms more personalised, while raising questions about how much control users will ultimately have over algorithmic recommendations and how the new preference signals will be used.

Key Takeaways

• New feature: Meta has introduced “Your Algorithm” on Facebook to help users customise recommendations.
• Where it works: The controls apply to Facebook Feed and Reels.
• Natural-language feedback: Users can describe topics they want to see more or less of in their own words.
• AI-powered recommendations: Large language models help interpret the context and intent behind written preferences.
• Real-time adjustments: Meta says recommendations can respond within the same session after users update their preferences.
• Simplified controls: The company has updated the three-dot menu to make feedback and personalisation options easier to find.
• Industry significance: The change gives users a more direct input into content discovery, but does not guarantee that every recommendation will follow their stated preferences.

Meta Launches “Your Algorithm” on Facebook

Facebook has traditionally relied on recommendation systems that infer what users might want to see based on their activity. Likes, comments, shares, viewing behaviour and interactions with pages and groups all provide signals that help determine which content appears in a person’s feed.

The new “Your Algorithm” feature adds a more direct way for users to communicate their interests. Instead of waiting for the system to learn from repeated behaviour, people can explicitly manage the topics they want Facebook to recommend.

For example, someone interested in technology news could ask to see more artificial intelligence developments and fewer celebrity updates. A person who enjoys cooking videos could request more recipes and less unrelated entertainment content.

Meta says users can find the dedicated controls through the three-dot menu. The company has also updated the menu on individual posts to make recommendation settings easier to access.

The feature is designed to give users more influence over recommendations without requiring them to understand how ranking algorithms work. It turns a traditionally indirect process—liking, watching or skipping content—into a more conversational exchange between the user and the platform.

However, the feature should not be interpreted as a complete replacement for Facebook’s existing recommendation system. Meta says written preferences will work alongside the signals its systems already use to personalise Feed and Reels.

How the AI-Powered Feed Customisation Works

The feature combines explicit user feedback with the information Meta’s systems already infer from behaviour. Users can indicate whether they are interested or not interested in a particular post and may be asked follow-up questions to explain their preference in greater detail.

These responses can be written in natural language, allowing users to provide more context than a simple reaction button offers.

Feature	What it does
Your Algorithm	Lets users add and manage topics they want to see in Feed and Reels
Interested	Signals that a user wants more content similar to a post or topic
Not interested	Signals that a user wants less of a particular type of content
Written feedback	Lets users explain their preferences in their own words
Follow-up questions	May help Meta understand what the user liked or disliked
AI-assisted ranking	Helps interpret the context and intent behind natural-language instructions

Source: Meta’s official announcement dated October 13, 2026.

Large language models are important because they can interpret context and intent expressed through ordinary language. A user may want more technology news but fewer promotional posts about technology products. A short written instruction can potentially communicate that distinction more clearly than a single like or dislike.

Meta says recommendations can respond in real time, including within the same session. That means changes to a user’s stated preferences may influence what appears next rather than requiring a long period of additional engagement before the system adjusts.

The company has not published detailed measurements showing how strongly these written instructions affect ranking compared with other signals. The exact influence of each input therefore remains unclear.

Why Meta Is Bringing AI Into Content Discovery

Meta operates social platforms where enormous volumes of posts, photos and videos compete for users’ attention. Recommendation systems help organise that content by estimating what each person is likely to find relevant.

The challenge is that engagement does not always reveal a user’s full intention. Watching a video does not necessarily mean the viewer wants to see more of the same topic. A person may watch something out of curiosity, dislike or a desire to understand a trend.

Explicit feedback can help address some of that ambiguity. If users can explain what they want, Meta may gain a clearer signal about their interests and preferences.

LLMs offer a way to interpret those explanations without requiring people to choose from a fixed list of categories. The system can potentially distinguish between related interests and understand more detailed requests.

For Meta, the change also fits its broader investment in artificial intelligence. The company has been incorporating AI into content creation, advertising, messaging and recommendation systems. Bringing natural-language interaction into feed personalisation extends that strategy to the process that determines what people discover on Facebook.

Still, better language understanding does not automatically guarantee better recommendations. The quality of the results will depend on how the system interprets instructions, balances competing signals and adapts to changing user interests.

What the Update Means for Facebook Users

The most immediate benefit is greater control over the topics that appear in Feed and Reels. Users who find their recommendations repetitive or irrelevant may be able to correct them more directly.

The tool could also help people with varied interests. Someone might want business news during the workday, sports content in the evening and cooking videos at weekends. Written instructions provide a way to communicate those preferences without relying entirely on past engagement.

However, the experience will still be shaped by Facebook’s broader ranking systems. Recommendations can come from sources a person follows as well as from accounts and pages they do not know. The new feature is intended to refine that experience, not necessarily eliminate unwanted content or guarantee that every recommendation matches a request.

Users should also distinguish between organic recommendations and paid advertising. Meta’s announcement focuses on shaping what people see in Feed and Reels. It does not establish that the new written preferences create new advertising-targeting controls or determine how advertisers can reach users.

The practical value of the feature will become clearer as more people use it and as Meta refines the system.

Meta Builds on Personalisation Tools Across Instagram and Threads

Facebook’s launch follows similar efforts across Meta’s other social platforms. Instagram introduced its own “Your Algorithm” controls earlier in 2026, allowing users to influence recommended content. Threads has also introduced a related feature called “Dear Algo,” through which users can express preferences about what appears in their recommendations.

The common idea is to make algorithmic personalisation more interactive. Instead of treating the recommendation system as a black box that users can influence only indirectly, the platforms are experimenting with tools that accept direct instructions.

These approaches may also help Meta make its recommendation systems more consistent across its services. However, each platform has different content formats and user behaviours, so the features may not operate identically.

The broader industry trend is significant. Social platforms, video services and online marketplaces increasingly use AI to determine which content, products or creators users discover. Giving people more direct control could become a competitive feature, particularly for users who feel that recommendations have become repetitive or overly focused on engagement.

Implications for Creators, Publishers and Brands

Changes to recommendation systems can affect creators and publishers whose reach depends on appearing in users’ feeds. If users can state their interests more precisely, content that clearly matches those interests may become more relevant to individual audiences.

For publishers, that makes topic clarity and audience relevance important. A technology publication that consistently produces useful AI explainers may appeal to people who explicitly request more AI news. A broad collection of unrelated posts may be less aligned with users who have refined their preferences.

However, Meta has not announced that the feature creates a new ranking guarantee for particular topics, keywords or publishers. Nor has it provided evidence that the update will increase organic reach for brands.

Advertisers should not assume that written preferences can be accessed as a new targeting category. The announcement concerns personalisation of Feed and Reels, and the company has not described the feature as an advertising product.

Creators should therefore continue to focus on original, relevant content and genuine audience engagement rather than attempting to optimise for an undocumented ranking signal.

Privacy and Transparency Questions Remain

The ability to state preferences in writing may help users control their feeds, but it also creates a more explicit record of their interests. A request for more financial news, health information or political content could reveal preferences that are more specific than ordinary engagement signals.

Meta’s announcement explains how written feedback can influence recommendations, but it does not fully detail how these statements are retained, how long their influence persists or whether the same data is used elsewhere within its systems.

Those questions matter because users should understand what happens to information they provide and how it affects their experience. Clear controls for reviewing, changing and removing preferences can help make personalisation more transparent.

Another question is how the system balances explicit requests against other signals. A user may ask to see more of a topic but continue interacting with different content. The way the platform resolves those conflicting signals will influence how much practical control the feature provides.

Meta’s rollout will need to demonstrate that conversational personalisation improves relevance without making the system difficult to understand or manage.

The Bigger Picture

Meta’s “Your Algorithm” feature reflects a broader change in how social platforms approach content discovery. AI is being used not only to recommend content but also to interpret direct instructions from users about what they want to see. By combining written preferences with behavioural signals, Meta aims to make Facebook’s Feed and Reels more responsive to individual interests.

The real test will be whether users experience a meaningful improvement in relevance and control. The feature could help reduce unwanted recommendations and make discovery more intentional, but its impact will depend on how strongly the system follows user instructions, how transparent the controls are and how Meta handles the preference data it collects. The announcement marks a product change, not proof that every user will receive a better feed.

Looking Ahead

Meta is likely to refine its personalisation tools as it gathers feedback on how people use written instructions and topic controls. The key indicators will be whether users can easily discover the feature, whether recommendations respond as expected and whether people return to adjust their preferences over time. More detailed information about rollout availability, ranking behaviour and data handling would help users and publishers better understand the change.

For creators and businesses, the safest approach is to keep producing content that serves a clear audience need rather than assuming the new feature creates a shortcut to reach. If more users begin actively shaping their feeds, matching content to specific interests could become increasingly important. But until Meta provides more evidence about the feature’s effects, its impact on reach, engagement and advertising should be treated as an open question.

Frequently Asked Questions

1. What is Meta’s “Your Algorithm” feature?

“Your Algorithm” is a Facebook feature that lets users add and manage topics they want to see more or less often in Feed and Reels.

2. How do users customise their Facebook feed?

Users can access the controls through the three-dot menu, select “Interested” or “Not interested,” and provide written feedback about the topics they want to see.

3. How does AI help customise Facebook recommendations?

Meta says large language models help interpret the context and intent behind users’ written preferences, allowing recommendation systems to adjust more directly.

4. Does the feature change recommendations immediately?

Meta says recommendations can update in real time, including within the same session. The company has not published detailed measurements of the effect for every user or type of request.

5. Is “Your Algorithm” available on Instagram and Threads?

Instagram has a related “Your Algorithm” feature, while Threads has introduced “Dear Algo.” The tools reflect a broader effort across Meta’s platforms to give users more influence over recommendations.

6. Will the feature change how Facebook ads are targeted?

Meta’s announcement focuses on organic recommendations in Feed and Reels. It does not establish that written preferences become a new advertising-targeting option.

7. What does the feature mean for content creators?

It may make clear audience relevance more important, because users can explicitly request topics they want to discover. However, Meta has not promised improved reach for any particular creator, publisher or subject.