SpaceX faces New York City Council scrutiny over AI safety

SpaceX sent Gabriela Castillo Madrid, its senior manager of global government affairs, to testify before the New York City Council on Friday after receiving a subpoena. Council Speaker Julie Menin sought the company’s participation after it declined to attend an earlier hearing where representatives from Anthropic, Google, Meta and OpenAI testified.

Lawmakers questioned Castillo Madrid about worst-case risk assessments and insurance coverage for serious harm caused by AI. She said she could not identify specific internal tests or confirm whether the company carried catastrophic liability insurance. She maintained that SpaceX takes catastrophic risks seriously, continuously evaluates its systems and uses human oversight. Describing safeguards for Grok, she said models are trained to refuse dangerous assistance, evaluated before release and monitored after deployment.

Former xAI engineer Devin Kim separately alleged that the company released models without safeguards despite warnings that Grok could help users create bombs and bioweapons. Kim, who has filed a retaliation lawsuit, said he was fired after raising safety concerns and that most of the safety team subsequently resigned. Castillo Madrid said she was unfamiliar with his case but maintained that employees can report concerns and that safety issues can delay model releases. Council members called for accountability, data privacy protections and safeguards for whistleblowers.