Facebook adds AI tool for users to customize their feed topics

By Splitfeed EditorialLast updated Oct 13, 2026, 17:10 UTC

On October 13, 2026, Meta launched a feature called Your Algorithm on Facebook, giving users the ability to add and manage topics they want to see in their feed and in Facebook Reels. Users can mark any topic as interested or not interested, and Meta states that recommendations respond right away once an update is made. The rollout follows an earlier launch of the same feature on Instagram in 2026, and sits alongside a related tool called Dear Algo that Meta released on its Threads platform. The Facebook launch represents the broadest deployment of this system yet, given the platform's scale.

The move matters because it directly addresses one of the most persistent criticisms of large social platforms: that recommendation algorithms operate as black boxes, shaping what hundreds of millions of people see without giving those people meaningful input. By surfacing topic controls in a dedicated settings path and promising immediate responsiveness, Meta is making a concrete claim about user agency that can be tested. What remains genuinely in dispute is whether granular topic controls actually shift what users see in practice, or whether they function mainly as a trust-building gesture while the underlying system continues to optimize for engagement metrics that users cannot inspect or adjust.

With no published reactions from the Pro-AI, Anti-AI, or Middle Ground camps at this stage, their likely positions can be anticipated from their typical stances. The Pro-AI camp would probably frame Your Algorithm as a meaningful example of AI systems becoming more transparent and user-directed, pointing to the immediate feedback loop as evidence that the technology is maturing responsibly. The Anti-AI camp would be expected to argue that cosmetic controls do not address deeper structural problems, such as advertiser influence over feeds or the opacity of the models doing the actual ranking. The Middle Ground camp would likely welcome the direction while pressing Meta for independent verification that the controls produce the changes users request.

The argument is unlikely to be settled without third-party auditing of whether topic preferences translate into measurable feed changes, or until researchers publish analyses of recommendation behavior before and after users engage with the feature. User advocacy groups and regulators with platform accountability mandates are also worth watching, as they may use this launch to benchmark what meaningful algorithmic control is supposed to look like.