Meta “Reveals” the Inner Workings of its AI Algorithm

a person holding up a smartphone with pictures on it.

Meta has suddenly decided to let the public see behind the curtain for how its artificial intelligence algorithm decides what content people see.

For the most part, big social media companies are very secretive about the inner workings of their content algorithms, but not Meta all of a sudden.

The tech giant, owner of Facebook and Instagram among others, is claiming a commitment to transparency as the reason why it wants to let users know just how its internal AI ranks content on both Facebook and Instagram.

According to reporting by Engadget, despite the seemingly arbitrary nature of social feed rankings on Instagram and Facebook, the reality is a complex system that depends heavily on user feedback.

As Nick Clegg, President of Global Affairs at Meta explains, “These systems make it more likely that the posts you see are relevant and interesting to you,”

This same Meta executive wrote an article on Medium in 2021 in which he explained how content ranking is balanced between user feedback and algorithmic systems.

Now, presumably with the blessing of his bosses, he’s claiming to show users of Meta’s two social networks how they can control the content they see and how well it’s shared.

According to Clegg, “This is part of a wider ethos of openness, transparency, and accountability,”

Given Meta’s known track record of anything but transparency in some of its dealings with the public and lawmakers, it might be a good idea to take some of these claims with a grain of salt.

Nonetheless, Clegg is arguing that Meta believes AI to be a crucial subject for open discussion at Meta and that this means sharing the company’s use of the technology with the public.

He elaborates,

“Our AI systems predict how valuable a piece of content might be to you, so we can show it to you sooner. For example, sharing a post is often an indicator that you found that post to be interesting, so predicting that you will share a post is one factor our systems take into account. As you might imagine, no single prediction is a perfect gauge of whether a post is valuable to you. So we use a wide variety of predictions in combination to get as close as possible to the right content, including some based on behavior and some based on user feedback received through surveys,”

To underscore these statements, Meta has published a series of 22 system cards that explain how its two social media platforms, Instagram and Facebook, rank content, how their AI predicts content preferences and how users can customize what they see.

instagram controls to influence what you see on instagram screenshot.

In the case of Instagram, much of the AI’s ranking “decisions” are based on interaction with previous posts by viewers.

This applies to feed posts and to stories. These sorts of actions also affect what a viewer sees more of when they go to the Explore section of their Instagram page or when they go to Search. Based on this, new accounts are suggested and specific posts by certain account types as well.

Meta’s system cards also reveal how the company uses an assortment of input signals that influence the system’s decision-making before it starts recommending things to users.

One of these is the way in which a user interacts with content. Another is how people in a user’s personal network of contacts interact with that same content.

a screenshot of instagram with the text see how activity informs ai models and the content you see.

Clegg also explains that contrary to popular opinion, users can adjust what their feeds and social media content suggestions show them.

According to the executive, users can indicate to these social platforms if they’re “interested” or “not interested” in content suggestions. He also adds that Meta wants to make its “show more” and “show less” options more visible in its apps.

Another user control mechanism that Clegg mentions is the Feed Preferences section that people can access in their Facebook account or the Content Control Center that their Instagram accounts give them.

On the other hand, despite all the claims of a commitment to transparency and yearnings for openness by Meta and its executives, the measures detailed above might be more about pandering to specific regulatory pressures.

Engadget itself speculates that much of these measures described by Clegg have more to do with upcoming European Union rules about increased social media platform transparency for these services within Union borders.

More fundamentally, despite Meta’s claims of responsiveness to user feedback, many users can still attest to a persistent flood of unwanted content in our own feeds.

For many (this writer included) the little touch buttons and menu options for dismissing content might as well be inert for all the difference they make in reducing spam, scams, unwanted ads and irrelevant content “suggestions”.

8 Tools for Photographers

Check out these 8 essential tools to help you succeed as a professional photographer.

Includes limited-time discounts.

Learn more here
Shotkit Journalist, Writer & Reviewer

Stephan Jukic is a technology and photography journalist and experimental photographer who spends his time living in both Canada and Mexico. He loves cross-cultural street photo exploration and creating fine art photo compositions.

Leave a Comment



Enter your email to be sent
today's Welcome Gift:
19 Photography Tools

🔥 Popular Now:

Shotkit may earn a commission on affiliate links. Learn more.