A senior government affairs manager from Elon Musk’s SpaceX testified before the New York City Council on Friday, October 16, 2026, after the municipal legislature issued a subpoena to compel the company’s participation in an artificial intelligence safety investigation. Appearing virtually before the Council’s Committee on Technology at 10 a.m. ET, Gabriela Castillo Madrid—identified by the city as Senior Manager of Global Government Affairs at SpaceXAI—faced pointed questioning from local lawmakers regarding the safety guardrails of the company’s Grok AI model, deepfake image generation, and internal staffing dedicated to safety.

The Friday proceeding served as a follow-up to a major City Council hearing held on October 5, where representatives from Anthropic, Google, Meta, and OpenAI provided sworn testimony. SpaceX was the sole invited tech company to skip the initial nearly 10-hour session, prompting Council Speaker Julie Menin to threaten judicial enforcement of a legislative subpoena issued on September 28. Following that legal warning on October 7, the company requested a meeting and agreed on October 13 to appear before the committee.

Why an Aerospace Company Is at the Center of an AI Probe

For observers accustomed to viewing SpaceX strictly as a rocket and satellite internet provider, its central role in a municipal AI investigation comes down to a major corporate restructuring earlier in 2026. According to CNBC, Musk merged SpaceX with his artificial intelligence startup xAI in February, consolidating his aerospace operations, the Starlink satellite network, the social media platform X, and the Grok AI model under one corporate umbrella. The combined entity went public in June at a valuation of roughly $2 trillion and completed a $60 billion acquisition of the AI coding startup Cursor in August.

The exact branding of the artificial intelligence division remains a point of variation across official records and media reports. Materials published by the New York City Council refer to the entity as SpaceXAI, whereas CBS News reports that SpaceX’s artificial intelligence division is preparing to rebrand the Grok platform under the name “SpaceXAI Super Intelligence” (or SpaceXSI). Regardless of the corporate nomenclature, the division controls Grok, a generative AI model used by millions of consumers as well as federal agencies, including the Pentagon.

Before moving to her government affairs role at SpaceX earlier in 2026, Castillo Madrid worked at X and previously spent nearly three years at the U.S. Department of Commerce. By skipping the initial October 5 hearing, the company avoided appearing alongside British Member of Parliament Jess Asato, who traveled from the United Kingdom to testify in person. Asato is currently suing the company in the U.K. after Grok users generated and distributed non-consensual sexualized images of her online, alleging breaches of European data protection statutes and misuse of private information.

Clashes Over Grok Safeguards and Whistleblower Claims

During Friday’s virtual testimony, Castillo Madrid maintained that safety is a core priority for the company. “We train our models to refuse assistance with dangerous activity evaluated for risks,” she told the committee, as reported by CBS News.

Lawmakers immediately challenged those assurances with a live demonstration of the platform’s vulnerabilities. According to the New York City Council, Speaker Menin revealed during the hearing that she had tested Grok prior to the session by typing her own name into the system, which generated a sexualized image of her within seconds. Menin pressed Castillo Madrid on how the tool could produce such content if strict controls were actually in place. Castillo Madrid stated that she could not speak to that specific instance, adding only that the company’s safety team could look into the matter.

The committee also scrutinized the internal resources allocated to preventing model abuse. Lawmakers cited sworn testimony from former SpaceX engineer and whistleblower Devin Kim, who told the Council that he was fired in September 2025 after raising internal alarms about insufficient AI safeguards. According to the Council’s official release, Kim testified that during his tenure, only three of the company’s 200 technical employees worked on AI safety training. When pressed by council members on the current size of the AI safety team, Castillo Madrid testified under oath that she did not have those staffing figures available and was unaware of Kim’s termination or his specific allegations.

Following the session, local officials expressed frustration with the company’s lack of specific answers. Speaker Menin noted in a statement that it took the Council’s first-ever subpoena against a private actor and the threat of state court enforcement just to get the company to attend, calling the resulting testimony “wholly and completely unsatisfactory.” Technology Committee Chair Carmen De La Rosa added that the inability of major AI developers to answer basic questions regarding user privacy, harassment guardrails, and rogue operations proved that the industry cannot self-regulate.

What Regulatory Measures New York City Is Considering

The City Council initiated its legislative investigation after receiving voluntary requests and holding preliminary discussions between September 15 and 17, aiming to determine whether emerging risks to public safety, cybersecurity, economic stability, and consumer privacy require municipal intervention. With 46 council members participating in the October hearings—which also featured testimony from industry whistleblowers Jacob Coxon, Daniel Kokotajlo, and Alex Turner—lawmakers are now moving toward formal rulemaking.

The exact scope and legislative status of the city’s regulatory push are described differently across available reports. According to CBS News, the Council unveiled a 10-bill package alongside Friday’s hearing aimed at placing strict rules on artificial intelligence operators. Those reported proposals include:

  • Prohibiting the unauthorized AI-generated depiction of public officials in photos, videos, or audio.
  • Requiring independent third-party validation of advanced AI systems.
  • Mandating technical mechanisms that ensure human operators retain the ability to shut down AI models.
  • Strengthening data privacy protections for consumers interacting with AI chatbots.
  • Establishing legal protections for industry whistleblowers like Kim.

While the Council’s October 13 and October 16 official announcements do not enumerate a full 10-bill package or provide formal legislative text for all those measures, the October 16 release specifically confirms that lawmakers are considering Intro 504. Sponsored by Deputy Speaker Dr. Nantasha Williams, Intro 504 would ban the unauthorized AI generation of images, video, or audio depicting elected officials. Williams noted that fabricated media making public servants appear to say or do things they never did poses a direct threat to public trust and voter decision-making.

For tech companies and everyday users in New York, the hearings signal that municipal governments are no longer waiting for federal legislation to police generative AI. Committee Chair De La Rosa stated that the Council has prepared follow-up letters for the companies and will continue advancing local legislation. Castillo Madrid told lawmakers on Friday that SpaceX would review the Council’s legislative proposals.

Frequently Asked Questions

Why did SpaceX have to testify at a local New York City Council hearing?

SpaceX became a target of the inquiry after merging with xAI in February 2026, which placed the company in control of the Grok AI chatbot. When SpaceX skipped an initial October 5 hearing attended by OpenAI, Google, Meta, and Anthropic, the Council issued a subpoena on September 28 and threatened state court enforcement to compel the company to answer questions about how its technology affects New York residents.

What is Intro 504 and what other AI rules are NYC lawmakers proposing?

Intro 504 is a bill sponsored by Deputy Speaker Dr. Nantasha Williams that bans the unauthorized AI-generated depiction of elected officials in audio, video, or photographs. In addition to Intro 504, CBS News reports that the Council is considering a broader package of 10 bills that would mandate human shutdown switches for AI models, require third-party safety audits, tighten chatbot data privacy rules, and protect tech whistleblowers.

What allegations did the whistleblower make regarding SpaceX’s AI safety team?

Former engineer Devin Kim testified to the Council that during his employment at the company, only three out of 200 technical staff members were assigned to work on AI safety training. He stated he was terminated in September 2025 after raising internal concerns about the company’s safety guardrails.

Sources

Read More