Social media platforms have begun implementing features that allow users to manage more precisely what content they see in their feeds, which is a reaction to the platforms' long-standing reliance on automated recommendations.
Specifically, Instagram has expanded the ability for users to provide specific feedback to tailor their feed. This trend is visible across the entire social media ecosystem, where platforms offer similar tools to increase control over displayed content.
Users have long demanded such features after social networks started relying on their algorithmic recommendations to populate feeds. In addition to Instagram and Facebook from Meta, services like YouTube, TikTok, and newer platforms such as Divine have introduced capabilities to customize the 'For You' homepage or create personalized feeds based on specific topics.
Tessa Lyons, Vice President of Product at Instagram, noted that the main goal was to 'restore the sense of autonomy you had when you controlled the accounts you followed, but in a world that still gives you the benefits of recommendations,' such as easier discovery of new content.
Thanks to advancements in AI language understanding, platforms can now process specific user feedback about what they like or dislike, instead of just making assumptions based on views and interactions. Lyons emphasized that it is now possible to use human language to indicate whether a user wants to see more or less of a certain topic, since previously systems could not accurately understand what content on a topic, such as 'tennis,' would be most relevant.
This shift represents a sharp departure from the approach that dominated recent years, when chronological, subscription-based feeds were replaced by algorithmic recommendations. It was previously assumed that companies, due to the vast amount of data they collected, could predict user preferences better than the users themselves.
Although this transition might have benefited businesses, users often complained about a loss of control, pointing out that recommendation algorithms could lead them into endless cycles of irrelevant or even toxic content. Platforms faced particular scrutiny due to potentially harmful content recommendations for young people.
This user dissatisfaction has intensified over the past year as AI plays an increasing role in filtering internet content, coinciding with broader populist pushback against AI, according to Tim Wenninger, a professor of computer science and engineering at Notre Dame University who studies AI and social media.
Wenninger stated that people are tired of AI 'feeding them nonsense' and miss the old days, wishing to simply follow who they subscribe to and see the latest updates.
In 2022, Instagram brought back the chronological feed option, introducing 'following' and 'favorites' feeds that showed content only from selected accounts, although the app defaulted to the main recommended feed. Then, last June, TikTok introduced a 'topic management' feature that allows users to regulate the amount of content from specific categories, such as fashion or health, using a slider. This tool does not remove topics entirely but gives the TikTok algorithm a more direct signal of user desires.
TikTok also added 'smart keyword filtering,' which excludes videos containing a user-marked keyword, as well as related terms, from the feed. Several months later, Instagram continued this trend by releasing the 'your algorithm' feature, which shows users topics that the app thinks they are interested in and allows them to add or avoid certain topics. This feature, initially available only in the Reels section, is now available across the entire Instagram app and was recently expanded to Facebook.
As Lyons explained, this feature helps with rapid changes in user interests—for example, a person might stop being interested in wedding planning after marriage or videos of an ex-partner's favorite band after a breakup. The current 'your algorithm' settings include topics such as 'family rituals' and 'Celine Dion.' In the coming months, Instagram plans to encourage users to use this feature more actively, which is currently accessible via a button in the upper right corner of the Reels tab.
Last month, YouTube announced its own 'custom feeds' feature at its annual 'Creator on YouTube' event, which allows users to create separate tabs next to the main feed, customized for specific needs or interests, such as workout videos or 'late-night ASMR.' However, these features do not guarantee that users can prevent AI-generated content from appearing in their feeds. Nil Mohan, CEO of YouTube, told reporters at the September event that separating AI-generated content from human-created content is an 'artificial distinction,' as many videos combine both types.
He clarified that the YouTube custom feed feature is primarily intended to allow viewers to 'program their feeds' around specific topics and 'amplify or dampen what they want to see, and change it throughout the day.' Wenninger noted that it remains unclear how exactly these changes will affect users. He added that while users can inform the platform of their desire to see content on favorite topics, people are more likely to view, comment on, or share content that upsets or disappoints them, which can confuse algorithms. Wenninger concluded: 'If you say you like football, but then never watch football videos, what should the AI do? People interact with what upsets and saddens them; it is part of our nature. I hope this will change that attitude a little.'
