Combating AI-Generated Deepfakes & Online Exploitation: Laws, Tech, and Support

The proliferation of AI-generated content, especially in the realm of explicit material, poses severe and multifaceted harms. Deepfakes, which use artificial intelligence to create realistic but fabricated videos or images, can be used to impersonate individuals, spread misinformation, and, most disturbingly, create non-consensual intimate imagery. Virtual child sexual abuse material (CSAM) refers to digitally created or manipulated images and videos depicting children in sexually explicit situations, which, despite not involving real children, can still inflict significant psychological damage on victims whose likeness is used and contribute to the normalization of child exploitation. Synthetic pornography, a broader category, encompasses AI-generated explicit content that may or may not involve identifiable individuals but contributes to a harmful online environment. The risks are particularly acute for children and teens, who are more susceptible to online grooming, exploitation, and the psychological distress caused by seeing themselves or others depicted in such material without consent. These forms of content can lead to reputational damage, emotional trauma, and even facilitate real-world harassment and abuse.

The International Centre for Missing & Exploited Children (ICMEC) plays a crucial role in monitoring and advocating for robust legal frameworks against child sexual abuse material (CSAM) worldwide. Since 2006, ICMEC has been tracking the legislative landscape concerning CSAM across the globe. Their most recent review indicates significant progress, with 156 out of 196 countries having introduced or enhanced legislation against CSAM. However, the depth of these laws varies considerably. While 111 countries meet four out of five essential legislative criteria, only 27 countries fully comply with all criteria, and a concerning 10 countries still lack any specific legislation. ICMEC's core criteria for effective CSAM legislation include: the existence of laws specifically addressing CSAM, a clear definition of CSAM, criminalization of technology-affiliated CSAM offenses, criminalization of the knowing possession of CSAM regardless of intent to distribute, and mandatory reporting of suspected CSAM by Internet Service Providers (ISPs) to law enforcement or designated agencies. These criteria are vital for creating a comprehensive and effective global response to the production and distribution of CSAM.

Beyond federal legislation, several US states have proactively enacted their own laws to address the growing threat of deepfakes and child sexual abuse material (CSAM). These state-level policies often carry significant penalties designed to deter perpetrators. For instance, in Tennessee, sharing deepfakes without permission is classified as a felony, with offenders facing up to 15 years in prison and fines of up to $10,000. Iowa has established the creation of CSAM as a felony, punishable by a maximum of five years in prison and a fine of $10,245 for a first offense. New Jersey's laws address the malicious creation and distribution of deepfakes, leading to potential prison sentences and fines up to $30,000. These state-specific laws demonstrate a commitment to tackling the issue at a local level, providing additional legal recourse and stronger deterrents against the creation and dissemination of harmful AI-generated content and CSAM.

Several key initiatives are at the forefront of combating deepfakes and CSAM through technological innovation and collaborative efforts. Safer by Thorn is a CSAM detection product that leverages machine learning to identify inappropriate content on digital devices, thereby helping companies reduce the risk of generating and hosting such material and aiding investigators in protecting children more swiftly. Another significant initiative is Lantern, launched by the Tech Coalition. Lantern is designed to detect harmful deepfake materials by utilizing signals associated with the perpetrators of CSAM. These signals can include email addresses, usernames, CSAM hashes, or keywords used in grooming and illicit transactions. Lantern is being integrated into the safety protocols of major tech companies such as Discord, Google, Meta, Roblox, Snap, and Twitch. The program aims to increase the reporting of criminal offenses to authorities and raise awareness about the predatory tactics employed by CSAM perpetrators, fostering a more secure online environment.

The emergence of AI-generated deepfakes, virtual CSAM, and synthetic pornography presents a formidable challenge to online safety and individual well-being. While the risks are significant, a multi-pronged approach involving international cooperation, robust national and state legislation, and advanced technological solutions is making headway. Initiatives like the TAKE IT DOWN Act, coupled with the diligent work of organizations like ICMEC and technological advancements from companies like Thorn and the Tech Coalition, are crucial in establishing legal deterrents and detection mechanisms. However, the fight against online exploitation is ongoing. Continuous vigilance, education, and the provision of accessible support resources for both youth and parents are paramount. By working together and leveraging the tools and knowledge available, we can strive to create a safer and more ethical digital future, free from the harms of AI-driven exploitation.