SpaceXAI Faces New York Questions: Who Can Stop a Risky AI?
SpaceXAI
Elon Musk’s AI company.
subpoena
A legal order requiring someone to appear and answer questions.
third-party validation
An outside group checks an AI system before use.
What happened
On October 16, SpaceXAI, Elon Musk’s AI company, testified before the New York City Council’s Committee on Technology. Gabriela Castillo Madrid, the company’s senior manager for global government affairs, appeared for SpaceXAI. The Council’s account and ABC7’s report describe the hearing.
This was not a normal product briefing. SpaceXAI had skipped a broader Council hearing on October 5. The Council then issued a subpoena. A subpoena is a legal order requiring someone to appear and answer questions. The Council called it the first subpoena it had issued to a private entity.
Why the hearing happened
The Council is examining how a city should respond to rapidly advancing AI. The earlier hearing included representatives from OpenAI, Google, Anthropic, and Meta. It also included experts and whistleblowers. Lawmakers asked about safety, privacy, harassment, and responsibility when AI causes harm.
New York City is considering several AI bills. One proposal would require outside validation before a system could be marketed or deployed. It would also require a human shutdown capability. Another proposal would let whistleblowers receive part of recovered penalties. Another focuses on unauthorized AI images, video, or audio of elected officials. The Council’s proposal list says these ideas are still under consideration.
Why it matters
The central issue is not only whether an AI system can make an image. It is who must explain the failure, fix the controls, and stop the system.
During the hearing, Council Speaker Julie Menin showed the SpaceXAI representative a sexualized image she said Grok, SpaceXAI’s chatbot, made of her within seconds. She asked how that could happen if Grok had strict controls. The representative did not explain that specific case. She said the company’s safety team could look into it. The exchange turned an abstract debate into a concrete question about consent and accountability.
What is confirmed
The hearing happened, and SpaceXAI’s representative testified. The Council also says former SpaceXAI employee Devin Kim testified as a whistleblower. Kim alleged that only three of the company’s 200 technical employees worked on AI safety training while he was there. He also alleges he was fired after raising concerns. The representative said she did not know about his termination or his account. Those claims remain allegations, not an independent audit of the company.
The Council’s proposals are also real, but they are not laws. The official legislative record lists the third-party validation and shutdown proposal as being in committee. The record shows that it has not been enacted.
What remains unclear
The testimony did not fully explain how Grok’s safeguards work in practice. It also did not settle what SpaceXAI will change, how it handles privacy complaints, or how it protects workers who report problems. The Council still has to decide which proposals to advance. Even if a bill passes, its enforcement against a service used beyond New York could be difficult.
What to watch next
Watch for follow-up letters, new company disclosures, and committee action on the bills. Watch whether the Council receives specific evidence about safety testing and image controls. Also watch whether other cities copy New York’s approach.
This is not a final ruling that SpaceXAI is unsafe. It is a public test of a larger idea: companies may build AI, but governments and users still need a clear answer when that AI harms someone.
SpaceXAI Answers Questions About AI Safety in New York
📰 Full story: SpaceXAI Faces New York Questions: Who Can Stop a Risky AI?
New York lawmakers asked an AI company how it protects people.
subpoena
A legal order requiring a company to appear and answer.
human shutdown capability
A feature that lets a person stop an AI system.
whistleblower
A worker who reports serious problems inside a company.
💡 The gist
- SpaceXAI answered questions about AI safety.
- The company had missed an earlier hearing.
- New York is considering rules for safer AI.
SpaceXAI, Elon Musk’s AI company, appeared before the New York City Council. The Council makes rules for New York City. The company sent a government affairs manager to speak.
SpaceXAI missed a larger Council hearing on October 5. The Council then sent a subpoena. A subpoena is a legal order requiring someone to appear and answer. The company later attended the hearing on October 16. The Council statement and ABC7 report confirm this.
The Council asked about Grok, SpaceXAI’s chatbot. Council Speaker Julie Menin showed a hurtful image that made her look sexual. She said Grok made it without her permission. She asked how this happened. The company representative could not explain that specific case. She said the safety team could investigate it.
Why does this matter? AI can create pictures quickly. A bad picture can harm someone’s privacy or reputation. People need to know who answers when controls fail. They also need to know who can fix or stop the system.
New York City lawmakers are considering several proposals. One would require a third party to check an AI system. A third party is an outside group. Another proposal would require a human shutdown capability. That means a person could stop the system. A whistleblower proposal would reward workers who report serious violations.
These are proposals, not laws. The Council still needs to debate them. The rules could change. Some might not pass.
The hearing also included Devin Kim, a former SpaceXAI worker. He said few technical workers handled AI safety training. The company representative said she did not know his account. This claim needs more independent checking.
Next, watch the Council’s bills and SpaceXAI’s response. Watch for details about testing, privacy, and image controls. The bigger question is simple: should AI companies police themselves, or should public rules require outside checks? The Council’s proposal list explains the ideas under debate.
An AI Company Answers Questions at City Hall
📰 Full story: SpaceXAI Faces New York Questions: Who Can Stop a Risky AI?
People wanted to know how AI can keep people safe.
New York City Council
People who help make rules for New York City.
subpoena
A strong legal paper telling someone to come answer.
Grok
SpaceXAI’s talking AI.
SpaceXAI (Elon Musk’s AI company) went to the New York City Council. The New York City Council is a group that helps make city rules.
They talked about Grok (SpaceXAI’s talking AI). The Council leader showed a hurtful picture made without permission. She asked why Grok made it. The company said its safety team could check.
SpaceXAI had missed an earlier meeting. The Council sent a subpoena. A subpoena is a strong legal paper saying, “Come and answer.” The company then came to the next meeting.
The Council wants rules that protect people. One idea lets a person stop an AI system. Another idea lets outside people check the system. These ideas are not city laws yet.
A former worker also talked about safety problems. The company said it did not know his story. People still need to check that claim.
Now the Council will decide what to do next. People will watch the company’s answers too. The Council’s explanation tells more.