Meta is altering how it communicates the factors that influence ad display to make the process more transparent. The company explained that it's updating a Facebook feature nearly a decade old called "Why am I seeing this ad?" in an effort to better inform consumers of how an AI technique called machine learning might have influenced the ad being served, among other factors.

Specifically, it will now highlight how user activity both on and off Facebook may inform its machine learning models, and it will detail how those models work to connect topics. The company for years has fought off rumors that it's listening to users' conversations in order to target them with relevant ads. In reality, the likelihood that Facebook is able to show ads that are relevant and creepy about brands or businesses users just discussed offline is due to Meta's highly developed ad technology that tracks users well beyond the Facebook app itself.

While effective targeting has been achieved in this manner, it has also given Facebook a somewhat creepy image problem, leading even the most rational of users to believe that Facebook has somehow found a way to listen to their conversation.

The social network says the changes to the "Why am I seeing this ad?" tool will help Facebook explain better to users how it is making those connections. It notes that it will now summarize information about ad targeting into topics that inform users how their Facebook activity and engagement outside of Facebook may have informed its machine learning models used in ad targeting.

Says Facebook in an announcement, this information could include things such as "liking a post on a friend's Facebook page or interacting with your favorite sports website." New examples and even illustrations detailing the workings of the machine learning models will also be provided.

The "Why am I seeing this ad?" feature itself can be accessed by clicking on the three-dot menu in the top-right-hand corner of any ad shown on Facebook. Once this update is added, users will see a screen explaining that there are a variety of factors, however, that go into determining how ads are both targeted and selected, as well as the specific actions surrounding this particular ad. For example, it may note that the advertiser targeted some particular demographic or that the user interacted with content related to the product of the advertiser.

It will also specify which Meta platform's activity was taken under consideration, such as engaging with other ads on the topic or posts and pages, or products displayed on Facebook Marketplace, or if a friend has interacted with a relevant page.

While some of that may not be surprising to users, further detail about off-Meta activity may be more surprising. Here, Meta clarifies that businesses and organizations share information with the company and that visiting other websites or using other apps and interacting with related products outside Facebook also contributes to ad targeting. Unfortunately, it doesn't appear to include a list of the websites or apps involved.

"The updates are founded on privacy researchers and policymakers whose comments were obtained, and have continuously reminded Meta to step up its game through heightened transparency in action: What does machine learning have in it for ads that show up? says Meta's Global Director of Ads and Monetization Privacy Pedro Pavón explaining the update. "Being transparent on how we use machine learning is necessary because it would assure that people are cognizant of the fact that this technology is a component of our ads system, and they know the nature of information it uses.

By making the case for greater transparency into how our machine learning models work to deliver ads, we hope to increase peoples' sense of security, he added, and heighten our accountability.

But alongside this new version of the ad tool, users will be sent to the settings that people can use to control what ads they want to see, and additional opt-out choices. There will be a link for these Ads Preferences settings, from other pages within the "Why am I seeing this ad?" tool.

Meta's refreshed ad tool reflects changes to the mobile-advertising ecosystem that have proved significant for the company. With Apple having recently rolled out its App Tracking Transparency (ATT) feature-a consumer privacy tool to prevent users from tracking, Meta has warned investors it will lose $10 billion of annual revenue. In this light, tracking data of the mobile users had given Meta better ways of personalizing their ads, and in effect, making them much more effective. As the number of consumers opting out rises, Meta has sought alternative data.

For example, according to CNBC, Shopify sellers were transferring their customer lists onto Meta's platforms to enhance the effectiveness of their ads.

Meta said in its Q4 earnings that there still were headwinds related to changes in ATT but was working to mitigate the impact through advertiser tools, ad formats that bring conversions on-site-like click-to-messaging-and longer-term AI investments. Revenue for Q4 was $32.17 billion, above estimates but marking the third straight quarter in decline and was down 4% year-over-year.

But the stock has popped because of Meta's promises of "a year of efficiency," its job cuts and its deemphasized focus on the metaverse in favor of AI work. The updated version of the ads tool will begin rolling out to global users on Facebook today and later this year to Instagram.