Facebook fails to promote specific ads to its target users, according to a North Carolina State University study.
Published in April 2022, the new research finds an average 30 percent inaccuracy of a user’s inferred interests with the ads targeted for them.
Using controlled experiments to obtain results, the authors created new accounts where they executed activities that will determine the accuracy of Facebook’s ad feature. They also invited 146 diverse participants from various locations to do the same.
Facebook’s ad preference manager identifies target users based on the content that they upload or like, the comments they post, the messages they send, and the pages they follow.
Aafaq Sabir, the research’s lead author and a doctorate student of the university, explained how Facebook’s algorithm can often take a keyword out of context. One example is a user mentioning a specific product. Even when the user dislikes that product, they would still get an ad similar to it.
Facebook often receives the brunt of criticism for its elusive algorithms which some countries consider to be divisive. In May 2022, it caught another controversy ahead of Australia’s media legislation.
The research, entitled, Analyzing the Impact and Accuracy of Facebook Activity on Facebook’s Ad-Interest Inference Process is co-authored by NC State Computer Science Assistant Professor Anupam Das and Computer Science major Evan Lafontaine.