Meta Introduces AI-Powered Tool That Lets Facebook Users Control Their Algorithms

Meta has introduced a new artificial intelligence feature called Your Algorithm, giving Facebook users more direct control over the content recommended to them. Announced on October 13, 2026, the tool allows users to tell Facebook which topics they want to see more frequently and which they would prefer to see less often. Instead of relying entirely on the platform’s existing recommendation system to interpret their behaviour, users can express their preferences in their own words. Meta says the feature is designed to make recommendations more relevant by allowing people to communicate directly with the technology that determines what appears in their Feed and Reels. The launch represents another step in the company’s efforts to integrate AI into the way people discover and interact with content across its social media platforms.

The new tool gives users several ways to adjust their recommendations. Through the three-dot menu on a post, users can access Your Algorithm or select options indicating that they are interested or not interested in particular content. They can also provide additional explanations about their preferences, helping Facebook understand why a particular topic appeals to them or why they want to see less of it. For example, someone who enjoys technology news but wants fewer celebrity updates can communicate those preferences directly. Meta says its AI-powered ranking systems will respond to these changes in real time, allowing recommendations to adjust within the same session rather than requiring users to wait for the system to learn from their activity over a longer period.

A key component of the feature is the use of large language models, or LLMs, to interpret natural-language instructions. Traditional recommendation systems often rely heavily on signals such as likes, comments, watch time, shares and browsing behaviour to predict what users might enjoy. Meta says AI models can add a more detailed understanding of context, intent and specific interests expressed in ordinary language. This could make recommendations more responsive to changing preferences, such as when a user temporarily wants more educational videos, business news, football updates or fashion content. Rather than simply recording whether someone interacted with a post, the system aims to interpret what that person wants to see and adjust its recommendations accordingly. However, the quality of the results will depend on how effectively the system interprets instructions and applies them to the large volume of content available on Facebook.

Your Algorithm also builds on Meta’s earlier efforts to introduce greater personalisation across its platforms. Instagram previously launched a similar feature, while Threads introduced a related tool called Dear Algo that lets users influence the topics appearing in their recommendations. Bringing the feature to Facebook extends this approach to one of the company’s largest social networks. The move comes as social media companies face increasing pressure to give users more influence over algorithmic feeds, which have traditionally been controlled largely by automated systems. More direct customisation could help users discover content that better matches their interests, although it does not necessarily mean that every recommendation will follow their instructions exactly. Facebook’s ranking systems still consider multiple signals when deciding which posts and videos to display.

The introduction of Your Algorithm could also have implications for creators, publishers, businesses and advertisers that depend on Facebook to reach audiences. If users begin adjusting their preferences more actively, content recommendations may increasingly reflect the topics people explicitly request rather than engagement patterns alone. Creators may need to focus on producing material that clearly serves particular interests, while publishers could benefit when audiences deliberately seek news and educational content. At the same time, the feature does not guarantee higher reach or visibility for any particular account, since recommendations remain subject to Facebook’s broader ranking systems and platform policies. The company has not said that users will gain complete control over every aspect of the algorithm. Instead, Meta is presenting the tool as an additional way to shape the content people encounter.

Meta says the new feature is rolling out on Facebook, with users able to access it through the updated three-dot menu and recommendation controls. The company has not indicated that the system will replace its existing ranking technology; rather, it adds a conversational layer that allows users to communicate their preferences more explicitly. The longer-term impact will depend on whether people find the controls easy to use and whether personalised recommendations become noticeably more relevant. If successful, the approach could influence how other platforms design their feeds, shifting social media personalisation towards systems that users can actively instruct instead of algorithms that primarily infer their interests. For Meta, the challenge will be to balance greater user control with the relevance, diversity and safety of the content shown across Facebook.