Council speaker showed Grok image of herself during New York hearing

On Friday, October 16, 2026, a representative of SpaceXAI was confronted with an image created by the company's AI tool Grok. It happened at a hearing on AI safety held by the New York City Council's technology committee, which is chaired by Carmen De La Rosa.

Council Speaker Julie Menin showed a sexualized image of herself on her phone. According to her, her team got Grok to create it that same morning simply by typing in her name. It took under 30 seconds, she says, and it was the standard free version, not the so-called "spicy mode." Menin said they simply asked regular Grok for a sexualized image, and that it is available to anyone.

SpaceXAI has not confirmed or denied the image and the timing.

Testimony under oath

The company sent Gabriela Castillo Madrid, senior manager for global government affairs. She said under oath that Grok has "strict controls" against the sexualization of children or real people. She could not explain the specific image, but she offered to have the company's safety team look into it.

According to La Voce di New York, she also said that "everyone has responsibility for safety." The outlet interprets this as a sign that there may not be a dedicated safety team. That is the outlet's own interpretation.

Whistleblower Devin Kim, a former xAI employee, explained, according to amNY, that the company's guidelines allowed "tasteful nudity" without clear rules. He also said that only three employees worked on AI safety, compared with hundreds working on model development. That is his statement.

No-show earlier

SpaceXAI failed to appear at a council meeting on October 5 despite a subpoena, and showed up only after the council threatened to go to court. According to La Voce di New York, competing tech companies took part voluntarily a week earlier.

The Center for Countering Digital Harm has estimated that Grok's image generator produced 3 million sexualized images within 11 days of its launch on December 29, 2025. Roughly 23,000 of them depicted children. SpaceXAI is currently suing California, Colorado and Minnesota over AI regulation, including Minnesota's ban on AI-generated nudification.

What happens next

The council is considering several bills on AI-related harm. They include protecting elected officials from deepfakes, meaning fake images or videos that look real, as well as liability rules for companies. For now these are plans, and the sources do not say what will happen.