Elon Musk Posts Creepy AI Video of Sydney Sweeney Generated by Grok

Elon Musk recently promoted new AI features on X by sharing an AI-generated video of actress Sydney Sweeney. the clip was created using X’s tool, Grok Imagine, and was meant to showcase improvements to the platform’s text-to-video technology.

In a post on X, Musk said Grok videos can now run up to 10 seconds and include significantly improved audio. The video features an AI version of Sweeney discussing those upgrades with a male voice, essentially explaining the new capabilities out loud. The prompt used to generate the video was submitted by user Alex Patrascu, who asked for a live-action style scene with cinematic camera work.

The post drew attention in part because it follows earlier controversial comments from Musk about Sweeney. In recent months, he shared other AI-generated videos of the actress and made remarks about her appearance at a film premiere, suggesting sympathy for her physical traits and captioning one post, “Can’t be easy.”

Grok Imagine launched in July 2025 and includes a “Spicy” mode that allows users to generate NSFW content. The tool has been used to create millions of sexualized images of women and minors, sparking international concern. California Attorney General Rob Bonta has opened an investigation into Grok over the creation and spread of nonconsensual sexually explicit material, calling it “shocking” and demanding immediate safeguards. At the same time, 35 state attorneys general are demanding stronger safeguards, and lawmakers in Alabama are proposing legislation specifically aimed at combating AI-generated sexual exploitation.

X has said it will restrict the generation of explicit images of real people in places where such content is illegal. However, users were still able to create that material even after the company’s announcement.

Regulators outside the U.S. are also watching closely. The European Union has opened its own investigation into X under the Digital Services Act, examining whether the company properly assessed and mitigated the risks of deploying Grok—particularly when it comes to illegal content and manipulated media.