A day before SpaceX’s initial public offering, which set stock market records, a giant inflatable figure of the company’s CEO, Elon Musk, appeared in Times Square in New York.

An unflattering caricature of a bare-chested Musk, with the words “SpaceX’s Grok makes AI child porn” on its chest and back, the inflatable was the centerpiece of a demonstration organized by the advocacy group Safe AI Now. The goal: tie the landmark financial offering to deepfake sexualized images of children generated by SpaceX’s AI platform, Grok.

The protest took place just outside Nasdaq’s global headquarters on West 42nd Street on Thursday.

A representative for SpaceX did not immediately respond to a request for comment.

A spokesperson for SAIN said in an email that because SpaceX owns Grok, it makes child porn. “A company that enables child porn is inherently unstable and puts American investors and retirement funds at risk. SpaceX shareholders are on the hook for every Grok lawsuit, criminal investigation, and regulatory fine that is coming,” the spokesperson said.

The organization describes itself on its website as “a coalition of faith leaders, family advocates, child development experts, online safety organizations, legal professionals, technologists, and concerned citizens working to ensure that artificial intelligence advances human flourishing.” SAIN is effectively anonymous; it does not identity any of its leadership or any individuals associated with the group on the website.

The effigy, the spokesperson said, was chosen as a metaphor for Musk and the companies he owns or is associated with, including the social media platform X and the satellite broadband provider Starlink, which have been absorbed into SpaceX along with Grok and xAI. (Musk’s automaker, Tesla, is separate.)

“Much like Musk and his companies, it is inflated, full of hot air, and could pop at any minute — it served as a warning to investors eager to buy into Musk’s SpaceX IPO today,” the spokesperson said.

Grok’s history of deepfakes

Ever since Musk introduced Grok in late 2023 and made it available to premium subscribers on X (formerly Twitter), the AI platform has had fewer guardrails than rivals such as ChatGPT and Claude.

It has a history of promoting antisemitism and hate speech while also allowing users, with its image-generation features, to do things such as undress photos of celebrities with AI-generated images or to create sexualized images of children. Those types of images have led to criminal investigations and lawsuits, and xAI made changes it said were meant to address Grok’s problems. 

But as Wired reported on Thursday, Grok continues to host sexualized deepfake images and videos of well-known women.