“These obscene posts put women, men, and children at extreme risk of harm,” three House Democrats wrote in a letter to tech billionaire Elon Musk over nonconsensual AI-generated sexual deepfakes last February.
Back in January, reports started coming in that X’s artificial intelligence bot, Grok, was complying with user requests to digitally alter other users’ photos to render them near-naked without their consent. Musk initially replied to these photos with laughing emojis. Now, he’s promised that his platform will not allow intimate deepfakes that sexualize users without their consent. Grok is not supposed to do this to people, he has said. If it is, it’s because of a “bug.”
Can we take Musk at his word? Given his history of online misogyny, it’s doubtful- and unfortunately, it also might not matter. Artificial intelligence is evolving exponentially, which is making it more and more difficult to design guardrails. And chatbots are notoriously sycophantic, which means that they want to make their users happy at any cost. They have encouraged delusional thinking, reinforced harmful user biases, and accepted conspiracy theories because of this. So, even if Musk and other developers do make efforts to stop intimate AI deepfakes, and the law works to curb it, user demand will probably keep it going.
And that’s because these bots are, in many ways, what men want women to be. So, even more so than the technology, we should blame the patriarchy.
From the stars to the streets
Nonconsensual sexual AI deepfakes occur when a person uses AI to remove most or all of a person’s clothes or put them in a sexual act. It’s a type of Non-Consensual Distribution of Intimate Images (NCII).
A few years ago, this issue mostly affected high-profile women because it was harder to generate deepfakes. Social media users began posting sexual deepfakes of Daisy Ridley, Emma Watson, and many other women actors as early as 2018.
That same year, after investigative journalist Rana Ayyub criticized India’s Bharatiya Janata Party (BJP) for supporting a child rapist, social media users bombarded her accounts with intimate deepfakes.
And if you haven’t heard of any of these cases, you’ve probably heard about what happened to Taylor Swift. In 2024, X users created sexual images of her at a football game using deepfake AI. These garnered 45 million views and thousands of reposts.
However, today, the technology is so easy to use that everyday women are at risk. Apple and Google stores host dozens of “nudify” apps with over 700 million downloads. Roughly 90 to 98% of all deepfake videos are these types of videos, and 99% feature women. LGBTQ+ people, people of color, and disabled people face increased risks as well. This reflects how, the further one’s body is from what Audre Lorde calls the “mythical norm” of a white, cishet, nondisabled man, the more entitled others feel to control that body. It’s also more common among minors than we’d like to think. Roughly 1 in 8 people ages 13 to 20 say they “personally know someone” impacted by these AI deepfakes, and it’s becoming an increasingly common bullying tactic in schools.
New weapons, old ideas
When women get justifiably angry about the fact that their bodies may, at any time, be altered, stripped, and shared like objects, a common response men give is, “Well, stop posting pictures on social media. It’s not that hard.”
There’s a lot wrong with this. First, even if the average woman might be able to keep her image off social media, many can’t. Politicians, activists, performers, and journalists, whose careers depend on public engagement, need to be on social media. It’s non-negotiable. Second, by telling women to hide their photos to avoid AI, these men are victim-blaming. Just like rape victims are blamed if they wear “revealing” clothing, women are blamed for deepfakes just for posting photos.
Like rape, this isn’t about lust- it’s about power. What happened to Rana Ayyub is an obvious (though horrifying) act of political “retribution,” as is the fact that 1 in 6 Congresswomen have been targeted by deepfakes. But consider also that X users generated Taylor Swift having sex in a football stadium after men got frustrated with her for being in this “masculine” space. And, returning to Emma Watson, the fact that she is now one of the most heavily targeted celebrities in AI deepfakes might have nothing to do with her being a prolific feminist, but it seems possible.
And you know what else this claim tells women? That they don’t belong in the public sphere of social media. Look back at the jobs that need it. Journalists. Politicians. Activists. Artists. These aren’t just careers. These are the people who talk about the way the world works and criticize it in a way that thousands will see.
Keeping it private
The renowned philosopher Jürgen Habermas first invented the concept of the “public sphere” to describe places where citizens from every background and status could debate on public issues. He believed this space was essential to create equality, state accountability, and democracy. Feminist theorists have discussed how, for hundreds of years, men have excluded women from this public sphere. (Because if a woman “stays in the kitchen,” she not only won’t vote, but also can’t access spaces where she could give her opinion on current events or learn from others.) This meant that their voices weren’t heard in public debate. It also meant that things like domestic violence and marital rape were seen as “private” issues, not symptoms of a misogynistic society.
In the 1960s, when women had the right to vote but did not have the right to obtain credit cards independently, serve on a jury, use contraceptives, or obtain a no-fault divorce, communities of women began to form “consciousness-raising groups.” These groups would discuss their lived experiences as women. From there, they gave names to those experiences and created language to explain male domination and forms of resistance. That’s why we have words like “patriarchy,” “rape culture,” “femicide,” and “objectification.”
Now, social media is our most important public sphere. And with the #MeToo movement– which began in 2017 as an online community of sexual assault survivors, and transformed into a campaign- we can see a digital parallel to the consciousness-raising groups of the 1960s. But it’s hard to know if AI deepfakes might affect future digital campaigns. A lot of women have spoken out online against deepfakes, which is a good sign. But studies show that 41% of women ages 15 to 29 self-censor to avoid online harassment. So, if nonconsensual sexual AI deepfakes become even more prominent, how much will (justified) fear stifle social media discourse- especially the type that criticizes misogyny?
Put a wrench in the (deepfake) machine
The good news is that there already are laws against deepfakes. The Take It Down Act, passed in May of 2025, criminalized non-consensual sexual AI deepfakes. The law also orders online platforms to create ‘request-and-removal’ systems, where the platform removes an AI deepfake within 48 hours of a victim’s request. However, to further secure survivors’ rights, we also must support passing the DEFIANCE Act, which would empower survivors to sue users and platforms for deepfakes without going to a criminal court.
And with a situation such as this, changing the law is not enough. Our culture must change, too. While federal law defines rape as a serious crime, survivors report only 16% of cases to police, and the justice system fails to hold nearly 98% of perpetrators accountable. Fear of disbelief and shame, and ignorance of their rights, often prevent survivors from reporting.
These thoughts happen because survivors have either internalized patriarchal messages that tell them rape is their own fault or because they fear other people might blame them instead of helping them. With AI deepfakes, we can see a similar phenomenon. Survivors experience psychological harm and serious (undeserved) damage to their reputations. And like rape survivors, survivors of AI deepfakes fear that reporting will only make things worse.
When shame and stigma enable violence, the best way people can resist it is by promoting other messages. In other words, be a safe person. Talk about this issue with people in your life to show women and girls that you support them, and to show men and boys that this behavior is unacceptable. These images, though terrifying, work by weaponizing misogynistic ideas. If we refuse to accept these ideas, more women can seek justice.
