Social media’s next evolution: user-controlled algorithms

Major social media platforms are rolling out tools that let users directly shape their recommendation algorithms, moving beyond passive feedback buttons toward explicit, AI-assisted preference controls. Threads, Instagram, and TikTok have each introduced features in recent months designed to give individuals more agency over what appears in their feeds. Threads launched a new "Your Algo" feature on July 16, 2026, building on its earlier "Dear Algo" tool [1]. The original tool required users to publish a public post, such as "Dear Algo, show me more posts about podcasts," to signal their interests [1]. The updated feature allows those same preference signals to be sent privately, and users can set a duration of one, three, or seven days for each request [1]. Instagram introduced its own control panel, called "Your Algorithm," which first appeared for the reels feed in December 2025 and expanded across the main feed and explore page in early June 2026 [1]. The tool surfaces the topics Instagram believes a user cares about most and provides toggles to see more or less of each subject [1]. Instagram head Adam Mosseri has stated that large language models now make it possible for recommendation systems to explain why content is shown and to accept explicit user instructions, a departure from older models that were opaque to the people using them [1]. TikTok’s "Manage Topics" tool, launched in 2024, uses a slider system to adjust the prominence of categories such as sports, travel, humor, current affairs, dance, and food in the "For You" feed [1]. In 2025, the platform expanded the tool with AI-powered Smart Keyword Filters that automatically block related terms; filtering out "remodeling" also suppresses synonyms like "renovation" and "renovations" [1]. These moves arrive as scrutiny of algorithmic amplification intensifies. Recommender systems, which use machine learning to analyze user behavior and personalize content, are the core mechanism behind feeds on major platforms and streaming services [4]. Critics have long argued that engagement-optimized algorithms favor content that provokes anger and outrage, contributing to political polarization and the spread of misinformation [3]. Research has also linked excessive social media use, driven in part by algorithmically curated feeds, to mental health problems, sleep disruption, and academic struggles [6]. A 2025 field experiment found that financial incentives to follow social media creators shifted participants’ policy positions and vote choices, with effects exceeding those of traditional campaign outreach, underscoring the persuasive power of algorithmically delivered content [9]. Meanwhile, audits of Data Download Packages from TikTok, Instagram, and YouTube have revealed inconsistent compliance with the GDPR’s Right of Access, including failures to disclose processing purposes and retention periods, raising questions about how transparent platforms truly are about the data feeding their recommendation engines [11].