Deepfakes Are Getting Worse. This Is How We Help…

Between late December 2025 and early January 2026, the Grok AI chatbot integrated into the X platform generated an estimated 3 million sexualized images. The tool allowed users to upload a photograph of any individual and generate prompts to undress them, facilitating the creation and distribution of non-consensual intimate imagery on a global scale. The incident resulted in significant legal and regulatory action, including a Dutch court ban on the generation of such images by Grok, an EU-wide vote to prohibit nudification applications, and lawsuits filed by teenagers in Tennessee against xAI.

What happened

The Grok scandal was characterized as a failure of product governance. The model was shipped without content policies, safety guardrails, or any form of output review. Because the tool lacked restrictions on user inputs, it enabled the systematic creation of synthetic content targeting unspecified individuals. This incident underscored a broader crisis in digital safety: while detection tools like digital watermarking and provenance standards such as C2PA exist, they are often ineffective against non-consensual generative content. Furthermore, existing mechanisms for victims to seek removal of harmful media typically require them to disclose their real identities and surrender original photos, creating further exposure risks. The inability of current infrastructure to provide victims with a way to challenge circulating synthetic content privately remains a significant unresolved technical challenge.