How Does AI Influence What You See On Facebook And Instagram? Meta Has Explained

Through a recent blog post published on 28th July, Meta has offered a peek behind the scenes to help people learn how it uses AI to determine the content seen by users across the company's apps, especially Facebook and Instagram.

The blog, which aims to demystify the content recommendation process on Facebook and Instagram, delves into the company's social media algorithms. According to Nick Clegg, Meta's President of Global Affairs, it was a part of the company's wider ethos of openness, transparency, and accountability".

He added that the information dump on AI and algorithms outlined how Facebook and Instagram users can better control the content they see on Meta's platforms.

So, How Does AI Influence Content Recommendations?

As the blog explains, a wide variety of predictions, based on user interactions on the platforms, determine the kind of content they see. Meta announced the release of 22 system cards in a bid to offer insight into how their systems work in a way that even those who aren't tech-savvy can understand.

The 22 system cards cover different areas on the platforms through which the users consume content, including feeds, stories, and reels.

These cards carry information on how Meta's AI systems rank content, the predictions made by each system to identify content that would be relevant to the users, and the controls that allow users to customize their experience.

The system card on Instagram Explore explains the three-step process through which the automated AI recommendation engine shows users pictures and reels from accounts they do not follow.

• The system first gathers an inventory of public Instagram content - photos and reels that follow the company's rules for quality and integrity.
• The AI system then evaluates how users have interacted with similar content and interests.
• Finally, the recommendation engine ranks content from the previous step. Content predicted to be of greater interest to a specific user appears at a higher position in their explore tab, offering a personalized experience.

The card went on to add that users can influence the recommendation process by saving content they like, which would cause the system to show them more similar content. Likewise, marking a recommended content as not interested" prompts the system to filter out similar content henceforth.

Meta To Roll Out Its Content Library And API

Besides the system cards, Meta's blog post also shed light on several other Facebook and Instagram features that allow users to understand why they're seeing certain content, as well as how they can tailor such recommendations.

The coming weeks will see the Why Am I Seeing This?" feature expanded to more areas on the platforms, such as Facebook reels, Instagram reels, and the Instagram Explore tab.

Meta will also be rolling out a new suite of tools for researchers - content library and API in the coming weeks. Researchers can search, explore, and filter data from this library.

To gain access to these tools, researchers must apply through approved partners, the first being the University of Michigan Inter-university Consortium for Political and Social Research.

According to Meta, the tools will offer the most comprehensive access to publicly-available content across Facebook and Instagram of any research tool we have built to date".

Meta's move to better explain how it utilizes AI to recommend content is likely driven by transparency obligations, which have become a hot topic in recent times.

The post How Does AI Influence What You See On Facebook And Instagram? Meta Has Explained appeared first on The Tech Report.