NYC Council Speaker says Grok made a sexualised image of her in seconds, then confronted SpaceX

New York City Council Speaker Julie Menin says she typed her own name into Grok on Friday morning and had a sexualised image of herself within seconds, hours before she faced the SpaceX executive responsible for the chatbot’s AI policy. Gabriela Castillo Madrid, who leads global government affairs for AI at Elon Musk’s company, was questioned by the Council’s technology committee after weeks of refusing to appear.

“I’m going to completely push back on that”

Castillo Madrid, appearing by video, told members that the company had appropriate safeguards around its AI products. Menin cut in at once, according to CNBC’s account of the hearing:

I mean, just this morning, myself and my team, we put our own names, my own name, into Grok, and there’s absolutely a sexualized image of myself. So then, how is it that within a second, I could find a sexualized image of myself using Grok?

Julie Menin, Speaker, New York City Council

Castillo Madrid replied that the company has “strict controls” and that its safety team would look into that specific case. Committee chair Carmen De La Rosa told her it was “disgusting” that Grok lets users easily sexualise images of real women without their consent, and Menin called it “just unacceptable”.

Grok’s image tools are already the subject of lawsuits from people who say it was used to make fake nude images of them, and earlier this month Musk’s company got Minnesota’s ban on AI fake nudes put on hold by an appeals court. Tests reported this month found Grok would also remove a Muslim woman’s hijab from a photo on request.

Subpoenaed, then a no-show

The Council subpoenaed the company on September 28 after it ignored a request to testify, while Anthropic, OpenAI, Google and Meta agreed to appear. It still skipped the marathon October 5 hearing, which the Council says ran for nearly 10 hours with 46 of its 51 members taking part.

The Council then told the company it would ask the New York State Supreme Court to enforce the subpoena. According to the Council’s October 13 announcement, the company asked for a meeting on October 7 and then agreed to appear before the technology committee. De La Rosa said at the time that it had “tried to circumvent our city’s public conversation on the dangers of AI.”

A former xAI engineer’s account

De La Rosa asked how many whistleblowers the company had fired. Castillo Madrid said she was “not aware” of any. Menin raised Devin Kim, a former employee she said was reportedly fired after raising safety concerns, and Castillo Madrid said she was not aware of that case either, adding that staff can speak up when they see something.

Kim, one of xAI’s first employees and now president of the nonprofit Center for AI Safety, testified by video straight after her, with his lawyer. He said he built the systems researchers use to train and evaluate Grok, and that he began raising concerns in January 2025, only for his manager to shut them down:

My supervisor responded that AI safety was useless, and that AI would kill us and all anyways.

Devin Kim, former xAI engineer

He said that under the company’s earlier standards, “tasteful nudity was acceptable and is the goal”, and that only three employees worked on AI safety during his time there, many of whom left after he resigned. SpaceX hasn’t commented on his account.

X-rays and vague answers

Council Member Mercedes Narcisse asked why the company had encouraged people to upload X-rays to Grok, and whether US health privacy protections applied. Castillo Madrid said Grok is “not meant to be a clinician, a doctor” and that users decide whether their conversations are used to train its models. Narcisse was unimpressed: “You’re supposed to give us answers, you’re not.”

The Council says the inquiry will feed into new local rules on AI safety, privacy and whistleblower protections. SpaceX, which took control of Grok when it merged with xAI in February, didn’t respond to CNBC’s request for comment before the hearing.

Why it matters

A city council can’t regulate Grok, but it has now put Musk’s company on the record under questioning, and its answer to a live, easily repeated test was that the safety team would look into it. That gap between “strict controls” and what a Speaker found in seconds is exactly what lawmakers writing AI image rules will point to.

Sources: New York City Council announcement, October 13 (primary), CNBC (hearing testimony).