Fake Journalists Created by AI for an Online Propaganda Campaign

Fake journalists created by AI

Middle East hot takes from “experts” who are actually fake personas were published on conservative sites like Washington Examiner. 

An investigation from the Daily Beast found out that fake journalists created by AI were used to publish pieces that were sharing an opinion favorable to certain Gulf states. The investigation concluded that at least 19 fake journalists, who were presented as analysts or political consultants, were used as authors of various editorial pages that were published in many conservative publications. Artificial intelligence was used to generate headshots of the fake authors who were used to make readers believe that the writers were real people. 

One of the fake journalists who shared an opinion on plenty of insights about the Middle East is Raphael Badani. He’s newsletter pieces were published in many conservative outlets, including Washington Examiner, American Thinker, or The National Interest. Yet, it turns out that Raphael Badani doesn’t exist and that his headshots were stolen from the blog of a San Diego startup founder. Badani is only one of the 19 fake personas that over the last year shared their opinion in more than 46 different publications. 

This isn’t the first time when artificial intelligence is used in such ways to leverage misinformation in the online environment. In fact, a 2019 report from The Associated Press discovered a fake profile on LinkedIn, a profile that was part of a network of likely spies who were trying to make connections with various professional targets. This time too, an AI was used to generate a fake headshot. 

AI-generated headshots can’t be traced back

The thing with AI-generated headshots is that they help build fake online personas with uniquely generated images that can’t be traced back to a source picture with a reverse image search. So, no one can prove that the persona is fake. 

Yet, the current generation of headshots of the fake journalists created by AI isn’t flawless. The profile pictures share a number of odd tells such as asymmetrical features, hair that blurs into nothing, or indistinct background imagery. 

According to the investigation from The Daily Beast, some of these odd features, plus others such as earlobes that are oddly melted or odd-looking teeth, were noticed in many of the headshots used by the 19 fake journalists. So, this suggests the fact that they are all part of the same thing, a coordinated campaign. 

The investigation also reveals that all fake journalists made the same mistakes which made them get caught, including presenting themselves as political consultants and journalists and lied about their academic and professional credentials. Plus, they all used fake or stolen avatars which were manipulated to can’t be traced back with reverse image search. 

Now, it isn’t clear who created the network of fake journalists. Yet, oddly enough, all pieces written by these fake writers share certain similar opinions, including that there must be more sanctions against Iran, they all praise certain Gulf states, and they all criticize Qatar. 

Twitter, the social media giant, also joined the initiative of exposing the network of fake journalists and suspended 15 accounts belonging to the counterfeit writers so far.