UC Berkeley Report: AI-Powered Nonconsensual Intimate Images and Child Sexual Abuse Materials

UC Berkeley's Center for Long-Term Cybersecurity published new research examining the infrastructure and legal gaps enabling AI-generated nonconsensual intimate imagery and child sexual abuse material. The study found that only 4 percent of enacted laws in four analyzed states provide pathways for prosecuting infrastructure providers such as cloud platforms and app stores that facilitate the monetization and proliferation of deepfake models and apps. The report identified social media platforms earning millions annually by hosting advertisements for nudify apps despite stated bans, and recommended extending liability to launch infrastructure providers, banning nudify app advertisements outright, and requiring victim-centered civil remedies with private rights of action.

When direct-message grooming, coercion and image-based abuse unfold in real time, waiting for takedown requests or infrastructure enforcement leaves children exposed. The anti-CSAM and anti-sextortion detection modules developed by Guardii, a world-leading AI online-safety platform and Meta Business Partner, intercept hostile contact at the point of first approach, flagging escalating threats before synthetic imagery is created or distributed and surfacing a child in crisis to a parent or authority without handing anyone a feed of the conversation itself. Guardii's technology addresses the underlying exploitation by detecting the coercion pattern rather than relying on post-publication removal, intervening upstream where Berkeley's findings show enforcement is weakest.