SpaceXAI will appear before the New York City Council after initially skipping a scheduled hearing, preventing a potential court dispute over its refusal to comply with a subpoena. The agreement gives council members an opportunity to question the company publicly about the risks associated with artificial intelligence.

The dispute follows a lengthy council hearing last Monday, during which AI industry representatives and former employees discussed potential dangers posed by increasingly advanced systems. Topics included the possibility of humans losing control over AI and the risk of catastrophic consequences. SpaceXAI, led by Elon Musk, did not participate despite being summoned to testify.

On Tuesday, Council Speaker Julie Menin said the council had warned SpaceXAI that it intended to seek judicial enforcement of the subpoena. The company subsequently contacted the council to arrange a meeting.

New York law provides several possible consequences for failing to obey a subpoena, including monetary penalties and measures to compel a witness to appear. Continued defiance can carry additional legal consequences, potentially including imprisonment.

Menin criticized the company’s absence and stressed the importance of public accountability for businesses developing powerful AI systems. She said: “Conspicuously absent was SpaceXAI, which is unacceptable. New Yorkers and the public at large deserve to hear directly from the companies developing this powerful technology,” Menin said. “Self-regulation is not enough, and I am glad that we have now secured SpaceXAI’s commitment to appear before the public in the coming week.”

Three former employees of leading AI research organizations testified at the hearing: Jacob Coxon, previously with Anthropic; Daniel Kokotajlo, formerly of OpenAI; and Alex Turner, formerly of Google DeepMind. Their testimony raised concerns about whether the industry can adequately manage the pace and potential consequences of AI development.

Coxon warned that humanity could eventually lose control of advanced AI systems. “On the current path … it is more likely than not that humanity loses control to these AIs, and it could end in human extinction,” Coxon said. Turner placed the odds of an “AI takeover” at approximately one in three.

Representatives from Meta, OpenAI, Anthropic and Google also faced questions from council members about how they evaluate and address severe AI-related threats. Menin has pointed to an exchange about how companies measure the possibility of the worst-case outcomes associated with the technology.

Morgan Dwyer of OpenAI responded: “I don’t know. I also don’t think it matters whether it’s 1%, 10%, or 20% chance.” She maintained that catastrophic risks should not be accepted, regardless of the estimated likelihood.

SpaceXAI is expected to appear before the council on Friday at 10 a.m. Council Member Carmen De La Rosa, who chairs the technology committee, said scrutiny of the company is especially significant because Musk’s technology businesses have an expanding presence in the federal government.

De La Rosa said the council plans to revisit questions raised during the earlier hearing while seeking more detailed answers about the company’s practices. She emphasized the need for information about personal data protection, safety measures and safeguards for employees who report concerns about their organizations.