Artificial Intelligence
New York City Council hears AI safety testimony
The New York City Council was due to hear from AI company representatives and former Anthropic researcher Jacob Coxon as it weighed guardrail proposals and a subpoena dispute involving SpaceXAI.
- The New York City Council was holding a hearing on artificial intelligence safety on Monday.
- Representatives from OpenAI, Meta, Anthropic and Google were set to testify as the council considered several guardrail proposals.
- Jacob Coxon, a former Anthropic researcher, was due to attend after posting that AI could pose risks to human lives.
- SpaceXAI declined to appear, and Julie Menin said the council planned to go to court over an alleged subpoena violation.
The New York City Council was due to hold a hearing on AI safety on Monday, with representatives from OpenAI, Meta, Anthropic and Google scheduled to appear. The council was considering several proposals for guardrails.
Council member Oswald Feliz, who chairs the Public Safety Committee, said there had been no guardrails in place. Julie Menin, the council speaker, said the idea of federal self-regulation did not make sense and said the council would urgently examine the risks.
Menin said one of the five companies called to testify had refused to attend: SpaceXAI. She said the council planned to go to court and argue that the company had breached a subpoena.
Jacob Coxon, a former Anthropic researcher, was also expected at the hearing after posting about his former employer and OpenAI. In those posts, he said AI could threaten human lives and prompted wider discussion among developers about the pace of development.
Anthropic's Logan Graham, OpenAI's Morgan Dwyer, Google's Alice Friend and Meta's Shane Cahill were set to testify before the council. Menin said the firms should answer questions and give their views on the proposed legislation under oath.
Named in this story
People
- Oswald Feliz
- council member and Public Safety Committee chair who raised guardrail concerns
- Julie Menin
- city council speaker who said the council would examine the risks and seek court action
- Jacob Coxon
- former Anthropic researcher due to attend after posting about AI risks
- Logan Graham
- Anthropic's Frontier Red Team head who was set to testify
- Morgan Dwyer
- OpenAI's head of policy development and operations who was set to testify
- Alice Friend
- Google's director of AI and emerging tech policy who was set to testify
- Shane Cahill
- Meta's AI policy director for legislation who was set to testify
Companies
- Anthropic
- AI developer whose former researcher was quoted and whose representative was due to testify
- OpenAI
- AI developer sending a representative to testify
- Meta
- AI developer sending a representative to testify
- AI developer sending a representative to testify
- SpaceXAI
- company that declined to attend and was said to be violating a subpoena
Governments and agencies
- New York City Council
- held the hearing and was weighing AI guardrail proposals
How the source tells it
The piece was a standard news report, but it repeatedly used officials' language of alarm and urgency around AI risk.
- alarm risk language was reinforced by repeated warnings about danger and a quote that AI could kill people
- urgency the council speaker framed the review as something to be done urgently
- public-duty framing the hearing was described in terms of owing the public answers and testimony