Tech

Facebook is working on deepfake spread and troll farms with the latest moderate push

Facebook has removed troll farms, false information spreaders, and deepfake image creators in its latest moderation activity.

The company’s latest Collaborative and Fraudulent Behavior (CIB) report (.PDF), released this week, lists Facebook’s latest efforts to reduce collaborative and fraudulent behavior across the network.

According to a March CIB report, Facebook investigated and wiped out “long-term” troll farms in Albania.

The members of the troll farm are primarily aimed at the Iranian audience and are believed to be associated with the political militant group Mojahedine Hulk (MEK), which consists of thousands of members.

MEK was expelled to Albania in the 1980s and now appears to operate a network of both real and fake accounts to disseminate information critical of the Iranian government and praising MEK’s activities. is.

According to Facebook, MEK-related content sharing surged through three separate clusters in 2017 and 2020, but most of the group’s efforts to grow its audience failed.

“Most of the accounts were run by Albanian operators who share the technical infrastructure on a daily basis,” the company said. “This meant that the same operator could run multiple accounts. Conversely, multiple operators could run the same account. These are physical locations shared by a collection of operators, the so-called Part of the characteristics of the Troll Farm. Computers and phones to jointly manage a pool of fake accounts as part of an influential operation. ”

In addition, Facebook is working on deepfake, an image generated by an artificial intelligence (AI) application. The company started removing fake images three years ago, but now the Generative Adversarial Network (GAN) is using deepfake to disguise itself as an independent news agency or research journalist.

After reviewing a survey provided by FireEye on the GAN networks in Spain and El Salvador, the company removed accounts and pages that published information about the mayoral elections at “spam-like” rates. Two more networks have also been wiped out, and Facebook’s total number of deletions is now seven operations using AI-generated images.

Social media giants have also documented the customary disruption of non-genuine networks. In March, a total of 14 CIB operations in countries such as Argentina, Egypt, Israel, Mexico and Georgia were suspended, removing over 1,100 accounts, 255 pages and 34 groups.

Facebook announced last month that it had detected and shut down China’s network of cyberattackers using the platform to distribute malware. Operators believed to be part of the Earth Empusa or Evil Eye group used fake profiles to target journalists and activists.

Previous and related coverage


Do you have any hints? Securely contact via WhatsApp | +447713 025 499, or key-based signal: charlie0


Facebook is working on deepfake spread and troll farms with the latest moderate push

Source link Facebook is working on deepfake spread and troll farms with the latest moderate push

Back to top button