Facebook is giving users a more active role in deciding what appears on their timelines, rolling out AI-powered tools that allow people to explain their content preferences directly instead of relying entirely on automated recommendations.
Announced on October 13, 2026, the update expands the platform’s “Your Algorithm” feature, allowing users to add or remove topics and guide the content recommended across Facebook Feed and Reels.
The new controls use artificial intelligence to interpret everyday language and adjust recommendations in response to users’ instructions. Facebook says these changes can take effect in real time, making it easier for people to tailor their experience as their interests evolve.
‘Your Algorithm’ Gives Users More Control
The expanded “Your Algorithm” feature provides a dedicated space where users can manage the subjects they want to see more or less frequently.
The tool is accessible through the three-dot menu on Facebook posts, which has also been updated to make personalisation options easier to find.
Rather than depending solely on the platform to work out their preferences from browsing behaviour, users can now state their interests directly and provide more specific guidance about the content they want to encounter.
The feature applies to both the main Feed and Reels, giving users a way to influence recommendations across two of Facebook’s major content surfaces.
New Feedback Tools Make Personalisation More Direct
Facebook is also expanding the feedback process attached to individual posts.
When users mark content as “Interested” or “Not interested,” the platform may ask follow-up questions to better understand what appealed to them or why they disliked a particular post.
Users can respond in their own words, explaining which topics they want to see more often or which ones they would prefer to avoid.
According to Meta, the additional information will help its recommendation systems refine their results immediately, including during the same browsing session.
The goal is to make personalisation more conversational, giving users a direct way to communicate their preferences instead of relying exclusively on likes, comments and viewing activity.
Large Language Models Help Interpret User Preferences
The update introduces large language models (LLMs) into the process of understanding and applying user feedback.
Facebook’s existing ranking systems already draw on numerous signals, including the posts people engage with and the videos they watch, to predict what might interest them.
With the new tools, AI can also interpret natural-language instructions, taking account of the interests, context and intent expressed in a user's request.
This means people can provide more detailed guidance about what they want to see, while the system combines those instructions with existing behavioural signals to shape recommendations.
Meta says this approach will make its ranking systems more flexible and responsive to individual preferences.
Will People Actually Change Their Settings?
Despite the additional controls, their impact will depend partly on whether users make use of them.
Facebook has offered various personalisation options for years, but many people continue using the platform’s default settings rather than manually adjusting their experience.
For example, users can switch to a chronological feed to view posts in time order. However, this does not become a permanent default, and Facebook returns to its algorithmically ranked experience when users leave and reopen the app.
Privacy settings provide another example of the gap between available controls and actual user behaviour. Even amid major public debates about personal data and online privacy, many users do not routinely review or change their preferences.
This presents a challenge for Meta: giving people more control does not automatically mean they will use it.
Meta Pushes Towards More Responsive Recommendations
The latest changes reflect a shift towards recommendation systems that respond not only to what users do but also to what they explicitly say they want.
By combining natural-language instructions with existing engagement data, Facebook is attempting to make Feed and Reels recommendations more closely aligned with users’ interests.
The redesigned controls may also make it easier for people to refine their experience without navigating through complicated settings menus.
Ultimately, the success of the feature will depend on whether users adopt it and whether the resulting recommendations better match their preferences. For now, Facebook is offering a more direct way to influence the algorithm rather than leaving content selection entirely to automated predictions.
Source: Meta