In the aftermath of the Cambridge Analytica scandal, Meta has taken steps to enhance transparency within its advertising network, notably through the Ads Library web portal and the API endpoint Ads Archive. Despite these initiatives, and potentially due to underutilization by the public and journalists, instances of exploitation endure on Facebook and Instagram. A recent illustration of this is the extensive RRN operation in 2023, involving hundreds of thousands of fake pages and advertisements 1 2.

This presentation aims to offer practical insights into navigating the API, highlighting current limitations and proposing effective workarounds. A key focus will be on Facebook’s categorization of “Social Issues, Elections or Politics.” Indeed, applying this label to an ad has two consequences : (1) heightened scrutiny by Meta moderation and (2) increased public accessibility to data such as cost and targeted audience. Deliberate mislabeling then becomes a standard modus operandi in foreign information manipulation and interference. This presentation will detail ongoing efforts to detect and quantify the prevalence of mislabeled ads using open semantic and network analysis libraries.