Facebook drops network of profiles that used photos generated by AI | Internet

The use of fake profiles to promote government advertising remains a headache for Facebook. Recently, the social network removed two networks accused of maintaining such practices, originating in China and the Philippines. The groups used artificial intelligence to generate profile pictures that attempted to mislead the platform’s identification.


To detect fake accounts, Facebook claims to focus the behavior of these profiles, in addition to their content.

Groups using a technique known as GANs (Generative Adversarial Networks) generally leave patterns of anomalies that can be easily detected. The practice has become common, according to the company specializing in social analysis Graphika.

Chinese group reportedly carried out activities related to US elections

The Facebook report revealed that a small part of the activities carried out by the identified group in China focused on the United States elections, with actions for and against Trump, in addition to another with support for the Biden-Harris duo. The Chinese network’s greatest interest, however, would have been in maritime security, with an emphasis on the South China Sea.

The pages generated by the group gathered 133 thousand followers on Facebook, in addition to 150 followers on Instagram. Advertising investments were relatively low, only $ 60, paid in yuan.

Fake post removed from Facebook

In the Philippines, the fake profile network consisted of several groups of connected activities. Facebook reveals that the accounts published content about local news, domestic politics, criticism of communism, young activists, among other subjects.

According to the social network, the activities were linked to the Philippine military and police. The report indicates that this network spent about $ 1,100 on advertising and gained more than 276,000 followers on Facebook, in addition to 5,500 followers on Instagram.

With information: Facebook and TechCrunch

Leave a Comment