Tall tales and fake faces
Between the AI-generated face, the nine year gap in activity, and the plethora of unlikely personal connections and overly convenient anecdotes, this is unlikely to be a legitimate account
Determining whether a given social media account is inauthentic is not always straightforward. Consider, for example, the case of X account @LovewinnLove, which commented on a variety of political topics between October 2023 and January 2024. This account, which sported an $8 blue verification checkmark alongside a StyleGAN-generated face, was created back in August 2014 but had no posts prior to October 2023, a gap of slightly over nine years. On their own these two attributes, while potentially suspicious, could have reasonable explanations. In combination, however, they suggest the possibility that an unused old account, possibly hijacked or purchased, was reactivated and redecorated to create this persona. (Since the account was created in 2014 and StyleGAN did not exist until 2018, it is impossible that the GAN-generated face is the account’s original profile image.)
There are multiple indicators that @LovewinnLove's "face" is GAN-generated, including the unrealistic teeth (particularly the visible portion of the bottom row), bizarre texturing and seams in the shirt fabric, and the telltale eye positioning characteristic of all StyleGAN-generated face images. As mentioned earlier, the use of an artificially generated face by a single social media account is not necessarily suspicious and does not on its own constitute proof that the account in question is inauthentic — perhaps the user wishes to remain anonymous, or simply finds a specific generated image aesthetically pleasing. In the case of this account, however, there are multiple additional indicators that all is not as it seems, including but not limited to the nine year gap between creation and first post.
A recurring theme of the @LovewinnLove account’s content is assertions that the account operator is a U.S. voter who has voted for Democratic candidates for their entire life, but is now switching parties due to the influence of “progressives” or “the left” on the Democratic party. In many posts, the account’s changing political views are bolstered by a mix of unverified personal anecdotes, claims of friends in high places, and tales of alleged involvement in world events, all of which conveniently support whatever point the account is trying to make. For instance, a story about having luggage stolen at a gas station in Oakland, CA is presented as evidence that sanctuary cities are mismanaged by Democrats, despite no link being demonstrated between this alleged crime and the city’s policies toward immigrants. Although some of @LovewinnLove’s assertions could be true individually, the sheer number of claims (a few of which are highly improbable), renders it extremely unlikely that all or even most of them are.
Let’s take a look at a few more examples. @LovewinnLove’s posts about the Israeli-Palestinian conflict are supported by claims to have a source in the IDF in a “decision making tactical role” and a mother in CENTCOM. Additionally, @LovewinnLove claims to have been involved in the Oslo peace process during the 1990s. Given that the account operator claimed to be in their 40s at present in other posts, this would make them a teenager when the negotiations began, making participation in high-stakes international diplomacy rather unlikely.
Similarly, @LovewinnLove’s posts about gender transition are backed by multiple dubious personal anecdotes. Supposedly the account operator has personally known "multiple kids who were pushed into puberty suppressants", and additionally has a minor child who was offered puberty blockers with no questions asked by any of the medical professionals involved (extremely unlikely).
The pattern continues with @LovewinnLove’s posts about their alleged career. Repeated claims to be a hiring manager are used to lend credibility to assertions that businesses will be hesitant to hire Harvard graduates. In other posts, however, the account operator claims to work in cybersecurity. While these two roles aren’t necessarily mutually exclusive, the job duties generally don’t overlap much, and both claims are unusually convenient in the conversations in which they occur.
In the same vein as the changing career descriptions, @LovewinnLove’s degree of familiarity with the Hebrew language and related back story elements vary from post to post. In one post, the account operator claims to require the assistance of a friend to translate Hebrew, in another they claim to be married to a Hebrew speaker, and on a third occasion they claim to both read and speak Hebrew and to have visited Israel personally many times. While there is some potential overlap between these explanations (perhaps the “friend” was also @LovewinnLove’s spouse), it is difficult to come up with a plausible scenario where @LovewinnLove is simultaneously able to read and write Hebrew due to multiple visits to Israel and requires a translator to understand the language.
One final inconsistency: in a series of posts on December 5th, 2023, @LovewinnLove repeatedly quoted one of X user @ashevat's posts with claims to know him. The specifics of this alleged relationship change from post to post, however — in some posts the connection is described as personal, in others professional. Notably, @ashevat did not interact with any of these posts, an interesting decision if the two parties were in fact acquainted and one party was attempting to draw attention to their relationship.
The research presented thus far in this article was posted in abbreviated form in a thread on X on January 24th, 2024. A few hours after this thread was posted, the operator of the @LovewinnLove account responded with a mix of mischaracterizations of the purpose of the thread and point-by-point attempts to explain the various anomalies highlighted therein.
The @LovewinnLove account’s initial replies to the thread were interesting for multiple reasons. For one, they repeatedly attempted to cast a purely analytical thread as something personal, both by characterizing the analysis as an “ad hominem” and by speculating on personal motives for writing it (“you have no other hobbies”, “creepy and stalky”, etc). Also, the “I actually forgot I opened that account in 2014” assertion is questionable — did they forget the account existed but remember the password? Are they using the same laptop or phone they used to set up the account over nine years ago (and did they somehow avoid getting logged out that whole time)? Possible, but not at all likely, and more reasonable alternate explanations such as the account being hijacked, purchased, or otherwise changing hands exist and are consistent with the presently available evidence.
The operator of the @LovewinnLove account attempted to provide explanations for several of the dubious claims in their posts: the mother in CENTCOM was intended as sarcasm (possible), the alleged role in the Oslo peace process was “enough to pass a polygraph test” (this doesn’t clarify anything), the changing explanations of Hebrew fluency are due to learning more via Duolingo (this is inconsistent with the claims to have picked up Hebrew while visiting Israel), and so on. While some of these explanations could be accurate individually, it is highly implausible that all of them are true (especially since no explanation is provided for some of the more outlandish assertions, such as the IDF source). Another noteworthy aspect of these replies is the repeated attempts to shift discussion away from the aspects of the account that don’t add up and toward debate of the account’s self-proclaimed political stances.
In a final wacky twist, the name of the account abruptly changed from @LovewinnLove to @MumbalaKumbala, and its posts became unsearchable. (A new account briefly surfaced with the @LovewinnLove handle, but has since been suspended by X.) Search results indicate that @MumbalaKumbala may have been the original name of the @LovewinnLove account, and that it changed names sometime on October 11th, 2023. Unfortunately, there are no posts archived under the @MumbalaKumbala handle on Wayback Machine, so this hypothesis regarding a potential October 2023 name change may be impossible to confirm.
Twitter says that MumbalaKumbala has 6,119 posts, but there's no tweets shown in the "Posts" or "Replies" tabs (but there's two tweets that are shown in the "Media" tab). Why would that happen? Or have you ran into any other accounts like that?