Meta Explains How AI Determines What You See on Instagram
Meta is letting people peek behind the curtain and learn how the company uses artificial intelligence (AI) to determine what people see when using Meta’s apps, including Facebook and Instagram.
As Engadget explains, social feeds can seem arbitrary, with no obvious explanation for why some users see one set of posts, while others see something else.
Meta says it is building on a prior commitment to deliver transparency to users and explains how user feedback ranks content across Facebook and Instagram. “These systems make it more likely that the posts you see are relevant and interesting to you,” says Nick Clegg, President of Global Affairs at Meta.
In 2021, Clegg wrote on Medium that “Content ranking is a dynamic partnership between people and algorithms.” Fast-forwarding to today, Clegg wants to make it clearer to Meta’s users how individuals can control what they see on Meta’s apps.
“This is part of a wider ethos of openness, transparency, and accountability,” promises Clegg.
Speaking about AI, Clegg acknowledges that while some people are excited about the possibilities ushered in by advancing AI technologies, others are concerned about the risks. From Meta’s perspective, Clegg says the company believes AI is critical but must be discussed openly.
“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,” Clegg says.
Meta has released 22 system cards for Facebook and Instagram that detail how each of the company’s AI systems ranks content, how the AI predicts content people will enjoy, and how users can customize their experience on Facebook and Instagram.
On Instagram, for example, a popular platform for photographers, Meta’s AI relies heavily upon interaction. Whether someone views and interacts with content on a Feed or Stories, these actions affect what that person sees in Explore, when using Search, and receiving suggested content directly on Instagram’s primary feed. AI also considers what content a user likes and suggests notifications about new accounts to follow, activity by connections, and more.
Meta also details various input signals that influence the system before it makes its recommendations. How a user interacts with content matters, but it also matters how people within someone’s social circle on the app interact with similar content.
It may often feel like there is no way to adjust how Meta’s AI delivers content to its users, but as Clegg points out, there are numerous ways that individuals can fine-tune content suggestions. On Instagram, people can now indicate whether they are “interested” or “not interested” in Reels, and Meta is working to make “show more” and “show less” more prominent within their apps.
“By using the tools available, you have the ability to shape your experiences on our apps so you see more of the content you want to see, and less of the content you don’t. To make this easier, we’ve created centralized places on Facebook and Instagram where you can customize controls that influence the content you see on each app. You can visit your Feed Preferences on Facebook and the Suggested Content Control Center on Instagram through the three-dot menu on relevant posts, as well as through Settings,” explains Clegg.
Meta is also expanding its “Why Am I Seeing This?” feature on Instagram and Facebook, allowing users to click on an individual piece of content and see how information from previous activity informed the machine learning model that underpins content suggestions.
Meta also explains that its Content Distribution Guidelines shape content suggestions, with the AI trying to filter out low-quality or harmful content. Meta says it cannot detail these signals, as doing so might make it easier for people to circumvent them.
Aside from that, Meta is relatively open about how it is using AI technology to serve content to its users.
As Engadget speculates, it may not be a coincidence that the European Union plans to enact the Digital Markets Act next year. The legislation will require online services like Instagram to be more transparent about the technologies in use and dictate that all platforms allow users to browse content with chronological feeds. Platforms must also ban ads targeting users based on religion, sexual orientation, ethnicity, or political affiliation.
Image credits: Meta
