行业动态
gizmodo.com ·
@Mike Pearl
加州州长 Newsom 签署两项获 AI 行业支持的 AI 监管法案
加州州长 Gavin Newsom 签署两项 AI 监管法案：SB 813 设立加州人工智能标准与安全委员会，负责制定一套自愿性 AI 安全标准；AB 1405 建立 AI 风险审计师注册制度，用于核验审计师资质。报道称 Anthropic 此前已支持这两项法案，OpenAI 也在签署前数小时宣布支持，OpenAI 公共事务负责人 Chris Lehane 表示公司将推动在能力评估、风险管理、人类控制以及何时放缓或停止开发等方面兼容的国际做法。Newsom 则呼吁联邦政府推出与当前紧迫性匹配的全国性监管。
AI推荐理由 报道加州州长签署两项 AI 监管法案：SB 813 设立加州 AI 标准与安全委员会制定自愿性安全标准，AB 1405 建立 AI 风险审计师注册制度；并给出 OpenAI 与 Anthropic 均表态支持的立场，以及 Newsom 呼吁联邦层面立法的表态。信息具体、可核验，但属政策新闻转述，对模型与开发者的直接技术价值有限。
72
正文
With the signing of two bills on Wednesday, California Governor Gavin Newsom is doing his part to crack down on AI after that spooky story about an AI researcher named Jacob Coxon quitting Anthropic and warning about how AI industry leaders think they might be hastening Armageddon. Oh, but it’s probably worth keeping in mind that the new California laws are well-liked by the AI industry.
As noted by Politico last month, Anthropic endorsed the bills. Anthropic’s chief rival, OpenAI, also announced its support for the bills in the hours before Newsom’s signing.
“We’ve reached a new chapter in AI capabilities, and that demands a new chapter for AI policy,” says the introduction of OpenAI’s statement, credited to Chris Lehane, OpenAI’s Chief of Public Affairs.
When asked by Politico about Coxon’s concerns about out-of-control AI in particular, Newsom said, “The concerns raised in recent incidents reinforce what California has long recognized: artificial intelligence holds extraordinary promise, but it must be developed and deployed with meaningful safeguards to protect the public.” The federal government, Newsom said, should “step forward with robust, national regulations that match the urgency of this moment.”
State Senator Jerry McNerney SB 813 establishes an organization called the California Artificial Intelligence Standards and Safety Commission. The commission will be tasked with creating a set of voluntary AI safety standards.
Meanwhile, Assemblymember Rebecca Bauer-Kahan’s AB 1405 creates an AI risk auditor registry—a way of checking auditors’ credentials.
For what it’s worth, the statement from OpenAI’s Lehane also says the company “will advocate for compatible international approaches to measuring capabilities, managing risk, preserving human control, and determining when and how development should slow or stop, even if that means slowing the advancement of model capabilities.”
Anthropic’s statement to Politico points to its history of creating safeguards. Such work, Anthropic says, “is also why we believe the world would benefit from the industry adopting a lawful, verifiable way to work together to pace how we release powerful models.”