Fake images, sexist posts target Harris, Congresswomen ahead of 2024 U.S. polls

In light of the recent allegations against American rapper Sean Diddy Combs’ surfacing in the lead up to the U.S. presidential bid, Democratic nominee and incumbent Vice President Kamala Harris is being targeted with disinformation using deepfakes and false narratives. As handles across the political spectrum make allegations against both the Democrats and Republicans’ ties to Combs, Harris has been subject to false claims that are sexist, misogynistic, and target her gender identity.

This recent targeting of the vice president lies at the center of a trend where female political leaders have been targeted globally on the basis of sex, themes of femininity and masculinity, transphobic commentary, and questioning their fortitude to hold positions of power. 

Fake posts on Harris run amok 

As AI-generated content is postulated to be a major threat this election cycle, former president and Republican candidate Donald Trump in a now-deleted post on TruthSocial, shared an edited photo of Harris alluding to whether she participated in Combs’ “freak offs.” 

Screenshot of the fake image being shared on social media. (Source: X/Modified by Logically Facts)

The real image, however, is from 2001 and it showed American television host Montel Williams with his daughter and Harris. Another handle with over 250,000 followers claimed that Harris was in a “sexual relationship” with Combs. The post amassed nearly 265,000 views at the time of writing this story. Several users also shared an edited picture in which Harris’ face was swapped with American stylist Misa Hylton, who is also Combs' ex-girlfriend, to further the narrative. 

Screenshots of the viral post circulating on social media. (Source: X/Facebook/Modified by Logically Facts)

In an X (formerly Twitter) thread, two verified handles exchanged sexist remarks targeting Harris. While one user with over 4,300 followers alluded the vice president exchanging sexual favors for a vote, another shared an edited image of Harris wearing lingerie with the caption, “Kamalasutra wif #Diddy at one of his freak-off parties (sic).”

Social media users shared unrelated and edited pictures calling Harris a “prostitute”, “call girl”, and a “sex worker”. (Source: X/Screenshots/Modified by Logically Facts)

The targeting of Harris was, however, not limited to her gender identity. Several posts also targeted her Indian roots. An X handle with just over 30 followers got over 12,000 views on a post of images of a bus stand in New York with advertisements allegedly showing Harris as a Hindu goddess standing over a boy with the text, “As a reminder Harris is 50% Indian 100% into cutting off children’s dicks (sic).” 

The images on the alleged advertisement showed Harris as Goddess Durga, with multiple arms, holding scissors in all of them. However, via crowdsourcing, Bellingcat’s Elliot Higgins was able to verify that the image was fake.

Commenting on the increasing use digitally manipulated images and videos, Center for Counter Digital Hate’s (CCDH) Research Head Callum Hood said, “Generally researchers or policy makers would like to look at something AI-generated and say that's obviously not real and it's being used to make more of a rhetorical political point. But research shows that people are quite poor at actually discerning whether an AI image is real or not. And that they do have an influence on people's beliefs even when they rationally understand that the images are not real.”

All in Together, a non-profit organization working towards equipping American women with civic education, found that public posts across platforms between July 29 and 11 August containing racist and/or sexist attacks against Harris had a possible reach of 6.62 billion views and impressions.

While disinformation targeting Harris’ gender identity appears to be significantly higher in proportion, other female politicians have also been subject to such narratives. Far-right Republican Congresswoman Marjorie Taylor Greene, who recently made headlines over spreading the conspiracy theory that the U.S. government can “control the weather” when referencing Hurricane Helene, was also targeted with sexist claims. 

An X post that now has a community note added to it falsely alleged Greene auditioned for American Idol with the claim “Shocking new footage of Rep. Marjorie Taylor Greene performing a raunchy act under a pseudonym has emerged online.” The February post got over five million views. 

Screenshot of the post with nearly five million views. (Source: X/Modified by Logically Facts)

However, the clip did not feature Greene and was debunked two years ago when it went viral. The same claim was shared on Instagram where a post got over 11,000 likes with comments under the post making comments that were also transphobic in nature. 

Similarly, handles shared edited images of Republican Congresswoman Elise Stefanik who has been referred to as Trump’s “cheerleader” and has been criticized for fueling false narratives on election integrity. Users shared edited images of Stefanik donned in Ku Klux Klan and Nazi outfits, possibly representing her pivot to a hard right stance from being considered a moderate. Examples of such posts can be found here and here. 

“We could see that Harris was already attracting quite a high level of abuse, but we also saw that on the other side among the Republicans individuals who have quite a lot of media coverage like Marjorie Taylor Greene receiving a lot of gender abuse as well,” Hood told Logically Facts. “One of the sort of common denominators of narratives is that they've received sexualised abuse. It can be based on how they look, suggesting that their political careers are a result of sexual favors, or direct threats of sexual assault.”

Tracking gendered misinformation over the years

A study published by CCDH in August 2024 analyzed 560,000 comments on Instagram posts of five Democratic and five Republican candidates and identified nearly 20,000 (1 in 25 comments) of them to be “toxic”. The study found that Instagram failed to act on 93 percent of abusive comments on women politicians and that most of these comments were based on their gender. 

“The cacophony of hate speech, threats, and gendered abuse we find flooding the comment sections of prominent women politicians is united in one shared purpose: to push women out of political life,” the study read. With the intersection of false claims, emerging narratives, and gendered commentary the lines between abusive behavior and disinformation often gets blurred. 

“You can get very high profile figures or alternative online news sometimes posting things that are at the junction of abuse and disinformation. So for example, once a high profile influencer starts suggesting that Kamala Harris got to where she was because of her racial identity or due to sexual favors - that has an influencing effect on other people keen to abuse her,” Hood said. “You can see that much of it is false, but it's also hateful.”

Over the last few years several eminent female political figures have been subject to such gendered targeting. Italian Prime Minister Georgia Meloni’s face was superimposed on a pornographic video and shared online, while former Finnish Prime Minister was vilified when a video of her dancing with her friends leaked. 

Similarly, a post with underlying transphobic messaging targeting former New Zealand Prime Minister Jacinda Ardern had over 430,000 views on X, while recently an AI-generated image of India’s Congress Party leader Sonia Gandhi smoking a cigarette went viral.

The impact of such claims, experts say, can result in discouraging women from entering politics, self-censoring, and a silencing effect on female politicians. Election cycles around the world have been marred by abuse and disinformation that disproportionately targets women—particularly, women of color and minority groups. 

In a report published by Center for Democracy and Technology, the non-profit organization concluded that while on an average Republican women candidates are more likely to be targeted with hate speech at 2.2 percent than Democratic women candidates at 1.4 percent, in this election cycle the effect is reversed and much higher for Democratic women of color who are subject to 19 times more hate speech than Republican women of color. It also found Harris was targeted with 40 percent less offensive speech compared to women of color candidates overall, and 45 percent less than African American women candidates.

While a conversation on the role of foreign interference in elections first began in the 2016 presidential race following platforms admitting it was weaponized, how much of foreign influence operations are gendered only recently began to be explored. 

A paper published in the International Journal of Communication found that X accounts from Russia’s military intelligence agency and the Internet Research Agency, Iran, and Venezuela spread gender-based disinformation designed to demobilize civil society activists, amplify government propaganda, and gain virality around divisive topics. 

With increasing sophistication in AI tools, and a record of social media platforms reportedly fueling the January 6 insurrection, how these giants maneuver content moderation is key. While Hood said Meta did respond to CCDH’s flagging of abusive comments under female politicians' posts, he also explained how it is difficult to monitor Facebook where such gendered disinformation is shared on closed groups. Similarly, Community Notes on X only feature after severe delays, and a recent report suggests TikTok failed to monitor harmful disinformation.

In response to the report, TikTok told Logically Facts that it was a small experiment that does not reflect its efforts to protect the platform in the 2024 election.  TikTok also said it prohibits misogyny, classifying it as a hateful ideology. 

Doctorate researcher at the Oxford Internet Institute, Alexandra Pavliuc said her work with researcher Nina Jankowicz and the Wilson Center led them to find three key narratives, “First, they [women] were targeted with sexualized narratives which claimed that they “slept their way” into their jobs or that they were too promiscuous to be trustworthy. Second is transphobic narratives, which claim that a successful woman is actually hiding that she is a man, therefore painting her as a liar and inciting transphobic sentiment. Third, is racist narratives, which we found went hand in hand with gendered ones for the black and south-east Asian women we studied.”

As experts argue that gendered disinformation is a national security issue, keeping in trend with how deepfakes disproportionately impact women; far-right AI-generated imagery of women being attacked by migrants, or of Harris being impregnated by Trump can affect perception. 

“People imagined a scenario of a flood of AI content that would be hard to discern or hard to disprove that could swing the election. But that was always the extreme end. I think under the surface actually, it's playing some role in the formation of people’s beliefs about the world, about elections, and about how they're going to vote,” Hood said.

(Disclaimer: Logically Facts has also reached out to X and Meta for a comment and the story will be updated if and when they respond.)