Background. As long as Facebook has existed, advertisements have been present in the application in one way or another. The ads have evolved and become more sophisticated over the years. Today, Facebook creates groups with members having specific attributes and advertisers requests groups for whom Facebook shows the advertisement. Besides this, Facebook receives information from other sources such as browser cookies and ad pixels. All information that Facebook receive or collect is used in their algorithms to target relevant advertisement for each user.

Objectives. To examine the possibility of Facebook eavesdropping through the microphone for marketing purposes and identify eventual keywords mapped between a spoken conversation and advertisement.

Methods. Five controlled experiments were performed with two test phones and two control phones. These were treated equally beside the test phones being exposed to spoken conversations containing randomly chosen products, companies and brands. The content of the phones was compared to see whether advertisement was adapted to the spoken conversation in the test phones but not in the control phones.

Results. No sponsored advertisements were present on the Facebook and Instagram application. Messenger contained ads indicating that Facebook might analyse the content of private messages to adapt advertisement. After adding the Wish application to the research, the results were still the same. Other contents in the Facebook news feed were analysed, however, the content analysed did not contain any evidence that Facebook eavesdrops on spoken conversations for marketing purpose.

Conclusions. The experiments conducted were not sufficient enough to trigger sponsored advertisement. Therefore, no indications were found that Facebook is eavesdropping through the microphone or not.