People spoke on Monday. They work with AI. They spoke at New York City Council.
AI業內人士週一在紐約市議會發言。
They warned that this new technology may hurt people.
他們警告,這項新技術可能傷害人類。
They asked for stronger oversight.
他們呼籲加強監督。
Local government is answering people's growing worries.
地方政府正在回應民眾越來越多的擔心。
Some workers recently left Anthropic, OpenAI, and Google DeepMind.
一些員工最近離開Anthropic、OpenAI和Google DeepMind。
They said at the meeting that these companies have a reckless culture inside.
他們在會上說,這些公司內部有魯莽的文化。
Former Anthropic engineer Jacob Coxon said these companies work with a startup mindset. That mindset is "move fast, break things." He said this is how they do things.
Anthropic前工程師雅各布·考克森說,這些公司用「快速行動、打破東西」的初創心態做事。
But he said this way does not fit the most powerful technology in history.
但他說,這套做法不適合史上最強大的技術。
He thinks there is no transparency and no safeguards. He thinks humans are more likely to lose control of AI.
他認為,如果沒有透明度和保障措施,人類更可能失去對AI的控制。
The worst case could even be human extinction.
最壞的情況甚至可能是人類滅絕。
The Trump government is not active.
川普政府態度被動。
Congress also does not act.
國會也沒有行動。
So state and city governments are thinking about controlling AI themselves.
所以州和市政府正考慮自己監管AI。
Trump once said fear of AI is a "hoax".
川普曾說,對AI的恐懼是「騙局」。
He signed an agreement. It lists steps companies can do by choice.
他簽署協議,列出公司可以自願做的步驟。
This basically lets companies control themselves.
這基本上讓企業自己監管自己。
Some leaders said AI development needs to slow down.
一些高管說過,AI發展需要放慢。
These leaders include Dario Amodei, Sam Altman, and Elon Musk.
這些高管包括達里奧·阿莫代伊、薩姆·奧爾特曼和伊隆·馬斯克。
But it is still hard for the whole industry to slow down together.
但整個行業一起放慢仍然有困難。
A group of workers also spoke at the hearing.
一個員工小組也在聽證會上作證。
This group speaks for OpenAI, Anthropic, Google, and Meta.
這個小組代表OpenAI、Anthropic、Google和Meta。
They said their systems are making daily life better.
他們說,他們的系統正在改善日常生活。
They also said they care about safety.
他們也強調他們重視安全。
City Council Speaker Julie Menin asked the company representatives about the worst disaster risk. She asked how big it is.
市議會議長朱莉·梅寧要求各公司代表說出最壞災難情境的風險有多大。
OpenAI's Morgan Dwyer answered that the chance does not matter. He said 1%, 10%, or 20% are all the same.
OpenAI的摩根·德懷爾回答,機率是1%、10%還是20%都不重要。
He said these levels are all unacceptable.
他說,這些水準都不可接受。
Menin criticized this answer as "reckless at best."
梅寧批評這個回答「最多算是輕率」。
The City Council is considering several AI rules.
市議會正考慮多項AI法規。
These proposals include information disclosure, third-party approval, and whistleblower rewards.
這些提案包括資訊披露、第三方批准和舉報人獎勵。
New York Governor Kathy Hochul has also taken action.
紐約州州長凱西·霍楚爾也已採取行動。
She is pushing for AI reporting and transparency standards.
她推動AI的報告和透明度標準。
Anthropic's Logan Graham said the company welcomes "smart regulation."
Anthropic的洛根·格雷厄姆則表示,該公司歡迎「明智的監管」。