Social Media's New Era: User-Commanded Algorithms

For many years, social media behemoths maintained significant control over the content users encountered in their feeds. While individuals possessed the ability to follow accounts, engage with posts, or suppress undesired content, the ultimate curation of what was displayed rested with sophisticated recommendation algorithms.

A notable shift is now underway, as social platforms are progressively empowering users to personalize these very algorithms, leveraging the capabilities of artificial intelligence.

Moving beyond rudimentary "Not Interested" functionalities, applications such as Threads, Instagram, and TikTok have begun rolling out innovative tools that enable individuals to actively train their own algorithms, thereby directly influencing the content that populates their feeds.

This evolving landscape signifies a fundamental transformation in the operation of recommendation systems. Social media feeds are transitioning from a passive, uniform "one-size-fits-all" broadcast model to something more akin to a personalized streaming service, where users can finely tune recommendations to align with their specific interests and exert greater autonomy over their viewing experience.

For users, the primary benefit of these customizable algorithms is a feed meticulously tailored to their individual preferences. Concurrently, for social media giants, this represents a strategic approach to amplify engagement by presenting content that users are most inclined to consume.

Let's delve into how social media platforms are granting users enhanced command over their recommendation algorithms.

On July 16, 2026, Threads introduced its new "Your Algo" feature, an advancement building upon the platform's "Dear Algo" tool, which first launched in February. The "Dear Algo" tool allowed users to influence their feed by publishing public posts, such as "Dear Algo, show me more posts about podcasts," to articulate their content preferences. With the advent of "Your Algo," users can now express these preferences privately, eliminating the need for public disclosure.

Adam Mosseri, head of Instagram, has previously stated that social media ranking models were historically constructed with technology lacking transparency for users. However, he notes that large language models (LLMs) are now capable of rendering recommendation systems more comprehensible by elucidating why specific content is displayed and enabling users to explicitly communicate their preferences.

Should users be uncertain about the types of videos encompassed within a particular category, an "information" button is available next to each topic. For instance, TikTok clarifies that its "Creative arts" topic includes diverse content such as "painting, drawing, graphic design, and art-related tutorials."