加州新出台的人工智能安全法案表明,监管与创新未必不可兼得。
内容总结:
【本报综合报道】美国加利福尼亚州近日签署生效的《第53号参议院法案》引发科技界对人工智能监管路径的深度讨论。该法案作为全美首个聚焦AI安全与透明度的立法,要求大型AI实验室公开其防范网络攻击、生物武器制造等灾难性风险的安全协议,并由州应急办公室监督执行。
青年倡导组织"编码人工智能"政策副总裁亚当·比伦指出,该法案印证了监管与创新并非对立关系。他在接受科技媒体采访时强调,现行立法完全可兼顾技术安全与产业发展,"企业已在实施法案要求的模型安全测试和模型卡片发布,但部分公司存在缩减安全投入的现象,这正是立法监管的必要性所在"。
针对业界担忧监管可能阻碍中美AI竞争的问题,比伦直言这种论调有失公允:"若将落后于中国归咎于《第53号法案》这样的基础性监管,实属避重就轻。"他认为产业界更应聚焦关键领域,例如支持国会通过《芯片安全法案》加强AI芯片对华出口管制,而非抵制各州针对深度伪造、算法歧视等具体风险制定的法规。
目前硅谷正通过政治行动委员会投入数亿美元影响州级选举,推动联邦层面出台取代各州立法的AI监管框架。比伦警示,若以狭隘的联邦立法架空各州监管权,将破坏美国科技治理的联邦制基石。值得注意的是,英伟达等芯片巨头出于商业利益反对出口管制,而特朗普政府近期亦出现政策反复,这些因素都使得AI监管之争更趋复杂。
《第53号法案》的落地过程表明,通过产业界与政策制定者的协商完全可能达成共识。尽管立法过程"混乱而曲折",但这种基于民主程序的治理机制,正是支撑美国科技经济体系的核心力量。
中文翻译:
加州州长加文·纽瑟姆本周签署生效的SB 53法案——这项以人工智能安全与透明度为核心的法案——证明了州级监管未必会阻碍AI技术发展。
在今日播出的《Equity》节目中,青年倡导组织Encode AI的公共政策副总裁亚当·比伦如此表示。他告诉TechCrunch:"政策制定者深知必须采取行动。从处理无数其他议题的经验中,他们明白立法完全可以既保护创新——这确实是我的关切——又能确保产品安全。"
本质上,SB 53是全美首个强制要求大型AI实验室公开安全协议细节的法案,特别聚焦如何防止模型引发灾难性风险,例如被用于攻击关键基础设施或开发生物武器。该法案同时规定企业必须遵守这些协议,并由应急服务办公室负责监管。
"法案要求的内容企业其实已在执行,"比伦指出,"他们进行模型安全测试,发布模型卡片。但某些企业是否开始在某些环节偷工减料?确实存在。这正是此类法案的意义所在。"
比伦特别提到部分AI企业在竞争压力下会放宽安全标准。例如OpenAI曾公开表示,若竞争对手发布缺乏相应防护的高风险系统,他们可能"调整"安全要求。比伦强调立法能督促企业履行既有的安全承诺,防止它们在竞争或财务压力下投机取巧。
相较于去年被州长否决的SB 1047法案,SB 53面临的公开反对声浪较弱。但硅谷和多数AI实验室仍普遍认为,几乎任何AI监管都会阻碍技术进步,最终拖慢美国在AI竞赛中超越中国的步伐。
正是这种认知促使Meta等公司、安德森·霍洛维茨等风投机构以及OpenAI总裁格雷格·布罗克曼等个人,共同向超级政治行动委员会注入数亿美元资金,在州级选举中支持亲AI的政客。这也解释了为何今年早些时候这些势力推动了一项试图禁止各州在十年内监管AI的暂停令。
Encode AI曾联合200多家组织成功阻击该提案,但比伦表示斗争尚未结束。倡导暂停令的参议员特德·克鲁兹正采用新策略实现联邦法优先于州法的目标。九月他提出《沙盒法案》,允许AI企业申请最长十年的联邦监管豁免。比伦还预判将出现旨在建立折中方案、实则架空州法的联邦AI标准法案。
他警告说范围过窄的联邦AI立法可能"为我们这个时代最重要的技术领域废除联邦制"。比伦坦言:"若说要用SB 53取代各州所有涉及AI及潜在风险的立法,我认为这绝非良策,因为该法案仅针对特定风险类别。"
尽管认同对华AI竞赛的重要性,也理解政策制定需要助推美国进步,但比伦强调扼杀主要针对深度伪造、透明度、算法歧视、儿童安全和政府应用AI的州级法案并非正确路径。"类似SB 53的法案会阻碍我们战胜中国吗?不会。声称这类法案会拖累竞赛进程,实在有违 intellectual honesty(学术诚信)。"
他补充道:"若真在意对华AI竞赛——我确实在意——就该推动国会实施出口管制,确保美国企业获得芯片。但行业诉求并非如此。"《芯片安全法》等提案旨在通过出口管制和追踪装置防止先进AI芯片流向中国,现行《芯片与科学法》也致力于提升本土芯片产能。但OpenAI和英伟达等科技巨头对这些措施的部分内容表现出抵触,理由涉及效能、竞争力和安全漏洞等考量。
英伟达确有苦衷——对华芯片销售历来占其全球营收重要份额,维持该市场符合其财务利益。比伦推测OpenAI在芯片出口议题上保持缄默,或许是为维系与英伟达等重要供应商的关系。
特朗普政府的矛盾态度更令局势复杂:2025年4月刚宣布扩大对华AI芯片出口禁令,三个月后却突然转向,允许英伟达和AMD向中国出售部分芯片以抽取15%收益。
"国会正推动《芯片安全法》等对华出口管制法案,"比伦指出,"与此同时,扼杀州级法案的舆论造势仍在持续——尽管这些法案监管力度其实相当温和。"
比伦强调SB 53展现了民主制度的实践——产业界与政策制定者通过协商达成共识。这个过程"虽然混乱曲折,但民主与联邦制程序正是我们国家和经济体系的基石,我期待我们能持续成功推进这种模式。"
他表示:"SB 53正是证明这套机制仍然有效的最佳例证。"
本文首发于10月1日。
英文来源:
SB 53, the AI safety and transparency bill that California Gov. Gavin Newsom signed into law this week, is proof that state regulation doesn’t have to hinder AI progress.
So says Adam Billen, vice president of public policy at youth-led advocacy group Encode AI, on today’s episode of Equity.
“The reality is that policy makers themselves know that we have to do something, and they know from working on a million other issues that there is a way to pass legislation that genuinely does protect innovation — which I do care about — while making sure that these products are safe,” Billen told TechCrunch.
At its core, SB 53 is a first-in-the-nation bill that requires large AI labs to be transparent about their safety and security protocols — specifically around how they prevent their models from catastrophic risks, like being used to commit cyberattacks on critical infrastructure or build bio-weapons. The law also mandates that companies stick to those protocols, which will be enforced by the Office of Emergency Services.
“Companies are already doing the stuff that we ask them to do in this bill,” Billen told TechCrunch. “They do safety testing on their models. They release model cards. Are they starting to skimp in some areas at some companies? Yes. And that’s why bills like this are important.”
Billen also noted that some AI firms have a policy around relaxing safety standards under competitive pressure. OpenAI, for example, has publicly stated that it may “adjust” its safety requirements if a rival AI lab releases a high-risk system without similar safeguards. Billen argues that policy can enforce companies’ existing safety promises, preventing them from cutting corners under competitive or financial pressure.
While public opposition to SB 53 was muted in comparison to its predecessor SB 1047, which Newsom vetoed last year, the rhetoric in Silicon Valley and among most AI labs has been that almost any AI regulation is anathema to progress and will ultimately hinder the U.S. in its race to beat China.
Join 10k+ tech and VC leaders for growth and connections at Disrupt 2025
Netflix, Box, a16z, ElevenLabs, Wayve, Hugging Face, Elad Gil, Vinod Khosla — just some of the 250+ heavy hitters leading 200+ sessions designed to deliver the insights that fuel startup growth and sharpen your edge. Don’t miss the 20th anniversary of TechCrunch, and a chance to learn from the top voices in tech. Grab your ticket before doors open to save up to $444.
Join 10k+ tech and VC leaders for growth and connections at Disrupt 2025
Netflix, Box, a16z, ElevenLabs, Wayve, Hugging Face, Elad Gil, Vinod Khosla — just some of the 250+ heavy hitters leading 200+ sessions designed to deliver the insights that fuel startup growth and sharpen your edge. Don’t miss a chance to learn from the top voices in tech. Grab your ticket before doors open to save up to $444.
It’s why companies like Meta, VCs like Andreessen Horowitz, and powerful individuals like OpenAI president Greg Brockman are collectively pumping hundreds of millions into super PACs to back pro-AI politicians in state elections. And it’s why those same forces earlier this year pushed for an AI moratorium that would have banned states from regulating AI for 10 years.
Encode AI ran a coalition of more than 200 organizations to work to strike down the proposal, but Billen says the fight isn’t over. Senator Ted Cruz, who championed the moratorium, is attempting a new strategy to achieve the same goal of federal preemption of state laws. In September, Cruz introduced the SANDBOX Act, which would allow AI companies to apply for waivers to temporarily bypass certain federal regulations for up to 10 years. Billen also anticipates a forthcoming bill establishing a federal AI standard that would be pitched as a middle-ground solution but would in reality override state laws.
He warned that narrowly scoped federal AI legislation could “delete federalism for the most important technology of our time.”
“If you told me SB 53 was the bill that would replace all the state bills on everything related to AI and all of the potential risks, I would tell you that’s probably not a very good idea and that this bill is designed for a particular subset of things,” Billen said.
While he agrees that the AI race with China matters, and that policymakers need to enact regulation that will support American progress, he says killing state bills — which mainly focus on deepfakes, transparency, algorithmic discrimination, children’s safety, and governmental use of AI — isn’t the way to go about doing that.
“Are bills like SB 53 the thing that will stop us from beating China? No,” he said. “I think it is just genuinely intellectually dishonest to say that that is the thing that will stop us in the race.”
He added: “If the thing you care about is beating China in the race on AI — and I do care about that — then the things you would push for are stuff like export controls in Congress,” Billen said. “You would make sure that American companies have the chips. But that’s not what the industry is pushing for.”
Legislative proposals like the Chip Security Act aim to prevent the diversion of advanced AI chips to China through export controls and tracking devices, and the existing CHIPS and Science Act seeks to boost domestic chip production. However, some major tech companies, including OpenAI and Nvidia, have expressed reluctance or opposition to certain aspects of these efforts, citing concerns about effectiveness, competitiveness, and security vulnerabilities.
Nvidia has its reasons — it has a strong financial incentive to continue selling chips to China, which has historically represented a significant portion of its global revenue. Billen speculated that OpenAI could hold back on chip export advocacy to stay in the good graces of crucial suppliers like Nvidia.
There’s also been inconsistent messaging from the Trump administration. Three months after expanding an export ban on advanced AI chips to China in April 2025, the administration reversed course, allowing Nvidia and AMD to sell some chips to China in exchange for 15% of the revenue.
“You see people on the Hill moving towards bills like the Chip Security Act that would put export controls on China,” Billen said. “In the meantime, there’s going to continue to be this propping up of the narrative to kill state bills that are actually quite light tough.”
Billen added that SB 53 is an example of democracy in action — of industry and policymakers working together to get to a version of a bill that everyone can agree on. It’s “very ugly and messy,” but “that process of democracy and federalism is the entire foundation of our country and our economic system, and I hope that we will keep doing that successfully.”
“I think SB 53 is one of the best proof points that that can still work,” he said.
This article was first published on October 1.
文章标题:加州新出台的人工智能安全法案表明,监管与创新未必不可兼得。
文章链接:https://qimuai.cn/?post=1352
本站文章均为原创,未经授权请勿用于任何商业用途