Hearing findings highlight need for common sense safety guardrails on artificial intelligence technology
NEW YORK – New York City Council Speaker Julie Menin and the Council’s Committee on Technology today held a follow-up AI safety hearing focused on Elon Musk’s SpaceXAI. The hearing took place following the Council’s first subpoena on record against a private actor, which was issued after SpaceXAI did not attend last week’s Council-wide hearing that included sworn testimony from Anthropic, Google, Meta, and OpenAI.
“It took a subpoena and the threat of continued legal action to compel SpaceXAI to show up and testify before New Yorkers today, and yet their testimony was simply unacceptable,” said Speaker Julie Menin. “Despite being under oath, their testimony raised far more questions than it answered and demonstrated the urgent need for governments around the world to take action and pass strong AI safety safeguards. Self-regulation is not enough, and the New York City Council is committed to ensuring continued innovation and growth while protecting the public with reasonable guardrails.”
The hearing included testimony from Gabriela Castillo Madrid, Senior Manager of Global Government Affairs at SpaceXAI. During the hearing, Speaker Menin told Castillo Madrid she created a sexualized image of herself within seconds using Grok. The Speaker asked how this could be possible if Grok had “strict controls” in place. Castillo Madrid was unable to speak to the images but said her safety team “could look into that specific instance.”
“The Council provided these major artificial intelligence companies an opportunity to build trust with New Yorkers interacting with their technology everyday, and they were unable to produce answers to basic questions around privacy protections for users, guardrails against harassment, whistleblower protections, and the risks associated with rogue operations of their technology,” said Council Member Carmen De La Rosa, Chair of the Committee on Technology. “It became clear to us that they cannot self-regulate. The Committee will continue to push conversations, with letters prepared for follow-up, and additional legislation and advocacy to rein in companies who do not believe they must answer to New Yorkers.”
The Council is considering legislation to place safety guardrails on artificial intelligence technology — including Intro 504, a bill sponsored by Deputy Speaker Dr. Nantasha Williams that would ban the unauthorized depiction of photos, videos, or audio of elected officials by artificial intelligence.
“As artificial intelligence becomes increasingly sophisticated, so does its ability to manufacture images, videos and audio that can make public officials appear to say or do things they never said or did. That poses a serious threat to public trust, particularly when New Yorkers are relying on information to make decisions about their government and the people who represent them,” said Deputy Speaker Dr. Nantasha Williams. “As I’ve worked to advance legislation addressing the risks of AI, including Introduction 504, my focus has been on establishing meaningful safeguards before the misuse of this technology becomes even more difficult to contain. Innovation cannot come at the expense of authenticity, and the public should not be left to navigate an increasingly convincing world of manufactured content without meaningful protections. As we examine the responsibilities of major AI companies, we have to be equally serious about the standards we expect them to uphold.”
At last week’s hearing, the Council heard testimony from British Member of Parliament Jess Asato; who is currently in a legal battle with SpaceXAI after Grok created nonconsensual explicit photos and videos using her likeness.
The hearing also included testimony from Devin Kim, a whistleblower at SpaceXAI who alleges he was terminated after raising safety concerns about Grok. He told the Council during his time at SpaceXAI only three of the company’s 200 technical employees worked on AI safety training. When asked whether she was aware of Kim’s termination and story, Castillo Madrid testified under oath that she was not.
Last week the Council introduced Intro-2605, a first-in-the-nation law sponsored by Speaker Menin. The bill would create a financial incentive for whistleblowers by allowing them to receive a portion of the fines or penalties recovered from artificial intelligence companies who break the law.
###