Whispers once confined to offline spaces have exploded online, manifesting as comments, mockery, shares, and screenshots. Despite a lack of consent, these actions have fueled widespread dissemination, reaching thousands within minutes and millions within hours. This is a daily reality for countless women and girls, as AI-generated deepfake content devastates lives, leaving justice elusive for most victims.
Deepfake harassment is a severe extension of broader gender-based digital violence targeting women and girls, a form of abuse that is escalating. While existing protective systems prove inadequate, the tools for causing harm are becoming increasingly cheaper, faster, and more accessible.
Understanding Deepfake Harassment and Its Prevalence
Deepfake technology involves manipulating images, audio, or videos using artificial intelligence (AI) to depict individuals saying or doing things they never did. Although the technology itself is not new, its weaponization against women and girls is a recent and rapidly spreading phenomenon.
- A 2023 report indicated that 98% of online deepfake videos are pornographic, with 99% of these depicting women.
- Deepfake video prevalence in 2023 was an estimated 550% higher than in 2019.
- Tools for creating this content are often free, widely available, and require minimal technical expertise.
- Once published, AI-generated content can be copied infinitely, saved to personal devices, and shared across platforms, making complete removal nearly impossible.
Victims’ Challenges in Reporting and Seeking Justice
Insufficient reporting stands as a significant barrier to accountability. For victims who do come forward, the justice system often becomes another source of trauma.
- Victims are frequently asked to repeatedly view and identify harassing content for police, lawyers, and platform moderators, often facing questions like, “Are you sure it’s not real?” or “Did you share private images before?”
- When a case reaches court, scrutiny often falls on the victim’s attire, relationships, and past behavior, rather than the perpetrator’s actions.
- A UN Women study found that 41% of women in public life experiencing digital violence reported subsequent offline attacks or harassment linked to it.
Why Deepfake Creators Rarely Face Legal Consequences
Despite the severity of the harm, prosecutions are rare, platforms frequently fail to act, and victims are often re-traumatized when seeking help. The reasons include:
Legislation Lags Behind: While less than half of countries have laws addressing online abuse, legislation specifically covering AI-generated deepfake content is even scarcer. Most “revenge porn” or image-based abuse laws predate deepfakes, leaving significant gaps. In many jurisdictions, deepfake pornography or AI-generated nudes fall into legal gray areas, leaving victims uncertain about whether the abuse is illegal and if perpetrators can be prosecuted.
Enforcement is Delayed: Even where laws exist, investigators require digital forensic expertise, cross-border coordination, and platform cooperation to build a case, yet most justice systems lack adequate resources for these needs. Evidence rapidly disappears as content spreads and copies multiply, perpetrators hide behind anonymity or operate across jurisdictions. Platforms are slow or unwilling to share data with law enforcement, especially in cross-border cases. Backlogs in digital forensics cause cases to stall before they even begin.
Tech Platforms Fail Victims: Companies have long hidden behind “intermediary” status to avoid liability for user-generated content.
The Path Forward: Essential Actions
While many nations and regions are taking steps, halting deepfake harassment demands urgent, coordinated action from governments, institutions, and tech platforms. Five key areas require immediate attention:
1. Laws Explicitly Addressing Deepfake Harassment
Governments must enact legislation that includes clear definitions of AI-generated abuse, prioritizes consent, imposes strict perpetrator liability, mandates swift content removal by platforms, and establishes cross-border enforcement protocols.
2. Justice Systems Capable of Investigation and Prosecution
Law enforcement agencies need training, resources, and specialized capacity for digital evidence collection and preservation. Digital forensic backlogs must be addressed, and international cooperation frameworks need to be swift, functional, and fit for purpose.
3. Accountable Platforms
Technology companies should be legally obligated to proactively monitor and remove abusive content within mandated timeframes, cooperate with law enforcement, and face tangible financial consequences for failing to act.
4. Meaningful Support for Victims
Free legal aid should be provided, alongside access to trained, trauma-informed law enforcement and legal professionals.
5. Education to Prevent Abuse
Training in digital literacy, consent education, online safety, and how to respond to abuse should begin at an early age and reach all segments of the population, as prevention is as crucial as prosecution.
UN Women emphasizes this is not a niche internet problem but a “global crisis.”
- In a recent high-profile case, British journalist Daisy Dixon encountered sexually explicit AI-generated images of herself on the X platform in December 2025, created using its Grok AI tool. The platform took days to geo-block the Grok function, allowing the harassment to continue spreading.
- In some cultural contexts, violated honor norms on digital platforms can serve as an online catalyst for “honor killings,” which may lead to extreme physical violence or death against women.
- According to a recent study, over half of deepfake victims in the United States have experienced suicidal ideation.
Meanwhile, several jurisdictions are beginning to act:
- Brazil amended its penal code in 2025 to increase penalties for psychological violence using AI or other technologies to alter women’s images or voices.
- The European Union‘s AI Act introduces transparency obligations related to deepfakes.
- The United Kingdom‘s Online Safety Act prohibits the sharing of digitally manipulated explicit images but does not cover the creation of deepfakes and may not apply if intent to cause distress cannot be proven.
- The United States‘ “Take It Down Act” explicitly covers AI-generated intimate images and requires platforms to remove them within 48 hours.
