Laws and Policies Against AI-Generated Deepfakes and Synthetic Pornography: What Governments and the UN Are Doing

AI-generated deepfakes, synthetic pornography, and virtual child sexual abuse material have become a pressing concern for policymakers, parents, and young people alike. Advances in generative AI make it possible to create realistic sexual images and videos of real people without their consent, and these materials can cause lasting psychological, reputational, and social harm. Understanding how governments and international bodies are responding is essential for anyone seeking to protect children and adults from this form of digital abuse.

Children and teens are among the most vulnerable to image-based sexual abuse. A fabricated image can be shared widely within a school or online community in moments, causing humiliation, anxiety, and depression. Virtual child sexual abuse material, even when no real child was photographed, normalizes exploitation and can be used to groom or extort victims. Because these harms are amplified by the speed and reach of social media, prevention and rapid removal are central goals of current policy.

The United Nations International Telecommunication Union (ITU) has encouraged social media companies to deploy digital tools that remove deepfake content. Its primary concerns include the effects of fabricated media on elections and the rise of financial fraud. Verification systems that authenticate whether images and videos are human-generated and real before they are shared are among the recommended solutions. The ITU stresses that a coordinated global approach is needed, since no single international organization currently focuses solely on detecting and removing manipulated harmful images. It is developing standards for video watermarking, which embeds a logo or unique code in a file to help identify its creator. This matters because video accounts for roughly 80% of internet traffic.

Since 2006, the International Centre for Missing & Exploited Children (ICMEC) has tracked global laws against child sexual abuse material. Its most recent review found that of 196 countries, 156 have introduced or strengthened CSAM legislation. Only 27 countries meet all of ICMEC's core criteria, and 111 meet four of the five. Ten countries have no CSAM legislation at all. The criteria examine whether national law specifically addresses CSAM, provides a precise definition, criminalizes technology-facilitated offenses, criminalizes knowing possession regardless of intent to distribute, and requires Internet Service Providers to report suspected CSAM to law enforcement or a designated agency. These gaps show that legal protection remains uneven worldwide.

The TAKE IT DOWN Act, signed into law on May 19, 2025, makes it a federal crime to publish non-consensual intimate imagery online, including realistic AI-generated pornographic images and videos depicting identifiable real people. The law also makes it unlawful to threaten to publish such imagery. Importantly, consent to the creation of an AI image does not equal consent to its online posting. The Act further requires social media companies to establish clear processes to remove this content within 48 hours of a victim's report, with companies given one year from the law's signing to fully implement these procedures.

Several U.S. states have enacted their own deepfake laws with significant penalties. Tennessee classifies sharing deepfakes without permission as a felony, with perpetrators facing up to 15 years in prison and fines of up to $10,000. Iowa treats the creation of CSAM as a felony, punishable by up to five years in prison and a $10,245 fine for a first offense. New Jersey imposes prison time and fines of up to $30,000 for creating and sending malicious deepfakes. These examples illustrate how states are tailoring penalties to the harms their residents face.

Technology plays a growing role in combating harmful synthetic media. Safer, a product from Thorn, uses machine learning models to identify potentially abusive and inappropriate content on digital devices. This helps companies reduce the risk of CSAM spreading on their platforms and enables policymakers and investigators to protect children more quickly. Detection tools like these complement legal measures by making harmful material easier to find and remove.

Launched by the Tech Coalition, Lantern is a cross-platform initiative that identifies and shares signals associated with CSAM perpetrators. These signals can include email addresses, usernames, CSAM hashes, and keywords used to groom victims or to buy and sell abusive material. Lantern is being integrated into the safety protocols of companies including Discord, Google, Meta, Roblox, Snap, and Twitch. The program also aims to increase reporting to authorities and raise awareness of the predatory tactics used by perpetrators.

Anyone affected by deepfakes should speak with a parent, a trusted adult, or a pediatrician. Parents seeking more information can consult the AAP resource Tips for Parents: Deepfakes, Synthetic Pornography, & Virtual Child Sexual Abuse Material. NoFiltr is an online community that offers young people resources and peer support for navigating sexual exploration in healthy and safe ways. It includes more than 500 advice submissions from teens about difficult online experiences, quizzes on digital safety topics such as sextortion, online grooming, and healthy online relationships, and guidance on seeking help. Individuals must be 13 or older to submit their own advice.

Addressing AI-generated deepfakes, synthetic pornography, and virtual CSAM requires coordination across international bodies, national legislatures, state governments, technology companies, and families. While laws such as the TAKE IT DOWN Act and detection tools such as Safer and Lantern represent meaningful progress, significant gaps remain in global legal protections. Continued policy development, technological innovation, and public education are all needed to protect children and adults from this evolving form of harm.