There’s precedent for regulating AI. Look to nuclear power.
This is read by an automated voice.
这是由自动语音读出的。
Please report any issues or inconsistencies here .
请在此处报告任何问题或不一致之处。
See more from the L.A.
查看更多来自洛杉矶的信息
Times in Google Search.
Google 搜索中的时间。
Set us as preferred
将我们设为首选
Barely a week passes without fresh evidence that artificial intelligence systems might spin out of control.
几乎一周过去了，就有新的证据表明人工智能系统可能会失控。
Public fears are rising.
公众的恐惧正在加剧。
Even the bosses of most big AI firms are calling for regulation and a slowdown.
甚至大多数大型人工智能公司的老板都在呼吁监管和放慢速度。
There’s little doubt that smart controls make sense, but the real question is how to design and implement them.
毫无疑问，智能控制是有意义的，但真正的问题是如何设计和实现它们。
For answers, AI might turn to another industry that faced similar challenges: nuclear power.
为了寻找答案，人工智能可能会转向另一个面临类似挑战的行业：核电。
The accident at Three Mile Island nuclear power station near Harrisburg, Pa., in 1979 was a wake-up call for the industry: Bad performance by any firm would destroy public confidence for all.
1979 年宾夕法尼亚州哈里斯堡附近的三哩岛核电站事故给该行业敲响了警钟：任何一家公司的糟糕表现都会摧毁公众对所有人的信心。
Many in the public were gripped by fear — made more palpable by the fact that Hollywood’s own tale of nuclear meltdown, “The China Syndrome,” opened in theaters just 12 days before the accident.
许多公众都被恐惧所笼罩，而好莱坞自己的核泄漏故事《中国综合症》就在事故发生前 12 天在影院上映，这一事实让这种恐惧变得更加明显。
Just as some AI firms are urging be done today, leading nuclear firms quickly swung into action and started regulating themselves — to make sure everyone in the industry avoided reckless behavior.
正如一些人工智能公司今天敦促采取的行动一样，领先的核电公司迅速采取行动并开始自我监管，以确保行业中的每个人都避免鲁莽行为。
They did this through extensive peer review.
他们通过广泛的同行评审做到了这一点。
Experts from each firm inspected each other , not just themselves.
各公司的专家互相检查，而不仅仅是他们自己。
To make the system work independently, they created a new organization, the Institute of Nuclear Power Operations.
为了使该系统独立运行，他们创建了一个新的组织：核电运行研究所。
The genius of the institute was that it didn’t confuse the two main roles for regulation.
该研究所的天才在于它没有混淆监管的两个主要角色。
One is learning about novel risks and helping firms, policymakers and the public learn how to avoid the worst ones.
一是了解新的风险，帮助企业、政策制定者和公众学习如何避免最糟糕的风险。
That’s what the institute does.
这就是该研究所所做的。
Over nearly five decades it has not only made the industry safer — and provided the model for a global version that helps review nearly every nuclear reactor on the planet — but also boosted economic performance.
近五年来，它不仅使该行业变得更加安全，并为全球版本提供了模型，帮助审查地球上几乎所有核反应堆，而且还提高了经济绩效。
(I am on its independent advisory board but not an employee.)
（我是其独立顾问委员会的成员，但不是员工。）
The other function of regulation is more punitive and restricting: the imposition of binding rules that constrain behavior and penalties on firms that stray.
监管的另一个功能更具惩罚性和限制性：施加具有约束力的规则，限制行为并对偏离轨道的公司进行处罚。
When today’s AI bosses make reasoned calls for “regulation” or “slowdown,” they envision a system for learning and pace — their version of the Institute of Nuclear Power Operations.
当今天的人工智能老板合理地呼吁“监管”或“放慢速度”时，他们设想了一个学习和节奏系统——他们的核电运行研究所版本。
But the big risk is that some firms — and, apparently, the Trump administration — so fear thickets of regulations and nonsensical constraints that they abhor any intervention.
但最大的风险是，一些公司——显然还有特朗普政府——非常害怕繁杂的监管和荒谬的限制，以至于厌恶任何干预。
Without all the big players on board, today’s calls for regulation will fail.
如果没有所有大公司的参与，今天的监管呼吁就会失败。
Self-regulation in the nuclear industry worked because the industry itself supplied the best experts.
核工业的自我监管之所以有效，是因为该行业本身提供了最好的专家。
Review by people who don’t know what they are looking for can be worse than nothing.
不知道自己在寻找什么的人进行审查可能比没有更糟糕。
For AI, these functions will need to be tailored to the risks and the ability to understand them.
对于人工智能来说，这些功能需要根据风险和理解风险的能力进行定制。
Open-source models might be different from closed ones, for example, and the risk analysis will need to be fluid as the technology changes.
例如，开源模型可能与封闭模型不同，并且随着技术的变化，风险分析需要不断变化。
It won’t be easy to rely on seconded employees because few firms will want to part with their key employees or let intellectual property slip away.
依靠借调员工并不容易，因为很少有公司愿意放弃关键员工或让知识产权溜走。
Luckily there are enough top engineers who are concerned about keeping AI aligned that they can staff a well-designed system for peer review — earning salaries good enough to stay engaged.
幸运的是，有足够多的顶级工程师关心保持人工智能的一致性，他们可以配备一个精心设计的系统进行同行评审——赚取足够高的薪水以保持参与。
In other industries that self-regulate — such as finance, biotechnology and aviation — firms are motivated less by altruism than by the fear of disaster and the knowledge that government might step in with more onerous regulation if private firms don’t step up.
在金融、生物技术和航空等其他自我监管行业中，企业的动机与其说是出于利他主义，不如说是出于对灾难的恐惧，以及知道如果私营企业不介入，政府可能会采取更严厉的监管。
AI’s biggest problem today might be that government seems to have lost that ability to intervene credibly, no matter who occupies the White House.
人工智能今天最大的问题可能是政府似乎已经失去了可信干预的能力，无论谁入主白宫。
The big AI firms will need to build an oversight consortium themselves.
大型人工智能公司需要自己建立一个监督联盟。
The good news is that in recent months the wisest calls for self-regulation are pointed in a similar direction.
好消息是，近几个月来最明智的自我监管呼吁也指向了类似的方向。
Last week Anthropic’s Dario Amodei called for “embedded evaluators” inside companies who make sure that risks are understood — a function similar to the nuclear power institute, but lacking needed independence.
上周，Anthropic 的达里奥·阿莫迪 (Dario Amodei) 呼吁在公司内部设立“嵌入式评估员”，以确保风险得到理解——这一职能类似于核电研究所，但缺乏必要的独立性。
In July, Demis Hassabis, Google’s AI czar, sketched a self-regulation system based on what is done in parts of the financial industry.
今年 7 月，谷歌人工智能沙皇德米斯·哈萨比斯 (Demis Hassabis) 根据金融行业部分领域的做法勾勒出了一个自我监管系统。
There’s a world of difference between saying AI needs regulation and designing a system that works.
说人工智能需要监管和设计一个有效的系统之间存在着天壤之别。
The concentration of the industry will help.
行业集中度将有所帮助。
Just a few firms that join a self-regulating consortium can control most of the frontier market.
只需加入自我监管联盟的少数公司就可以控制大部分前沿市场。
Just two countries, the United States and China, also dominate.
只有美国和中国这两个国家也占据主导地位。
Success probably requires starting with as small and focused a group as possible — big firms working inside the U.S. — and then expanding out.
成功可能需要从尽可能小且集中的团队（在美国境内运营的大公司）开始，然后向外扩张。
A clock is ticking for the AI industry.
人工智能行业的时钟正在滴答作响。
None of the AI bots slipping their chains, to date, is probably enough to inspire a serious self-regulatory system.
迄今为止，没有一个人工智能机器人能够摆脱束缚，可能足以激发严格的自我监管系统。
Fierce competition — evident in the huge capital expenditures of the industry — makes any regulation of the whole industry vulnerable to firms that drag their feet or refuse.
激烈的竞争——从该行业巨额资本支出中可见一斑——使得对整个行业的任何监管都容易受到那些拖延或拒绝的公司的影响。
Waiting longer, however, risks making regulation harder because there could be more new firms with new ideas and risks, not just here but also in China and other countries that learn how to operate at the AI frontier.
然而，等待更长时间可能会导致监管变得更加困难，因为可能会有更多拥有新想法和风险的新公司，不仅在美国，而且在中国和其他国家学习如何在人工智能前沿运营。
David G.
大卫·G.
Victor is a professor of innovation and public policy at UC San Diego and a nonresident senior fellow at the Brookings Institution.
维克多是加州大学圣地亚哥分校创新和公共政策教授，也是布鲁金斯学会非常驻高级研究员。
L.A.
洛杉矶
Times Insights delivers AI-generated analysis on Voices content to offer all points of view.
Times Insights 对 Voices 内容提供人工智能生成的分析，以提供所有观点。
Insights does not appear on any news articles.
见解不会出现在任何新闻文章中。
The following AI-generated content is powered by Perplexity.
以下人工智能生成的内容由 Perplexity 提供支持。
The Los Angeles Times editorial staff does not create or edit the content.
《洛杉矶时报》编辑人员不会创建或编辑内容。
The article argues that AI regulation should learn from nuclear power, where a major accident prompted firms to create an independent peer-review institution that improved safety without relying solely on punitive government rules.
文章认为，人工智能监管应该向核电学习，核电的一次重大事故促使企业创建一个独立的同行评审机构，以提高安全性，而不仅仅是依赖惩罚性的政府规则。
It contends that regulation has two distinct purposes: developing shared knowledge about emerging risks and imposing binding requirements with penalties.
它认为监管有两个不同的目的：发展有关新兴风险的共享知识和施加具有约束力的要求和处罚。
A specialized AI consortium could perform the first function, while government retains authority for the second.
专门的人工智能联盟可以执行第一个功能，而政府保留第二个功能的权力。
The piece suggests that peer review must be conducted by technically capable experts who understand what to examine.
这篇文章建议同行评审必须由技术能力强、了解要检查的内容的专家进行。
The system should account for differences between open-source and closed models and evolve as the technology changes.
该系统应该考虑开源模型和封闭模型之间的差异，并随着技术的变化而发展。
At the same time, the article maintains that self-regulation will work only if all major frontier-AI companies participate.
同时，文章认为，只有所有主要的前沿人工智能公司都参与进来，自我监管才能发挥作用。
The industry’s concentration makes it possible to begin with a focused U.S. consortium and expand internationally later.
该行业的集中度使其可以从专注于美国的财团开始，然后再向国际扩张。
The proposal is consistent with emerging policy efforts that emphasize independent or third-party assessments of frontier-model risks, including state measures requiring safety frameworks, audits and risk-mitigation procedures.
该提案与新兴政策努力一致，强调对前沿模型风险进行独立或第三方评估，包括要求安全框架、审计和风险缓解程序的国家措施。
[1]
[1]
The article warns that competitive pressure and the enormous cost of developing advanced models could encourage firms to delay, weaken or evade voluntary controls.
文章警告说，竞争压力和开发先进模型的巨大成本可能会鼓励企业推迟、削弱或逃避自愿控制。
It therefore calls for an oversight structure with genuine independence rather than evaluations conducted entirely within individual companies.
因此，它需要一个真正独立的监督结构，而不是完全在个别公司内部进行评估。
The principal opposing view is that industry self-regulation places rulemaking and enforcement in the hands of companies whose financial interests depend on rapidly expanding AI capabilities.
主要的反对观点是，行业自律将规则的制定和执行权交给了那些经济利益依赖于快速扩张的人工智能能力的公司。
A Brookings analysis argues that voluntary commitments are inadequate unless policymakers establish who writes the rules, verifies compliance, enforces violations and protects people harmed by AI.
布鲁金斯学会的分析认为，除非政策制定者确定谁来制定规则、验证合规性、强制执行违规行为并保护受人工智能伤害的人，否则自愿承诺是不够的。
[3]
[3]
From this perspective, embedded evaluators may not be independent if companies select, fund and control them.
从这个角度来看，如果公司选择、资助和控制嵌入式评估人员，他们可能并不独立。
Effective oversight would require publicly defined standards, clear authority to delay or block releases, and accountability beyond executives and investors.
有效的监督需要公开定义的标准、延迟或阻止发布的明确权力以及高管和投资者之外的责任。
[3]
[3]
Critics also reject the assumption that technical expertise alone can determine acceptable risk.
批评者还拒绝接受这样的假设：仅凭技术专业知识就可以确定可接受的风险。
Decisions about release schedules, safeguards, liability and victims’ rights involve public values and democratic accountability, not merely engineering judgment.
有关发布时间表、保障措施、责任和受害者权利的决策涉及公共价值观和民主问责制，而不仅仅是工程判断。
[3]
[3]
The article’s preference for beginning with a small group of major U.S. firms contrasts with concerns that a private consortium could entrench the dominance of the largest companies and leave smaller developers, affected communities and the public with little influence.
文章倾向于从一小部分美国大公司开始，这与担心私人财团可能巩固最大公司的主导地位并使规模较小的开发商、受影响的社区和公众几乎没有影响力的担忧形成鲜明对比。
A competing policy approach favors enforceable government standards alongside industry expertise.
竞争性政策方法有利于可执行的政府标准以及行业专业知识。
Current state initiatives include transparency requirements, protections against discriminatory automated decisions, independent audits and frontier-model risk frameworks, while no comprehensive federal AI bill had passed in the current session.
当前的州举措包括透明度要求、防止歧视性自动决策、独立审计和前沿模型风险框架，而本届会议尚未通过全面的联邦人工智能法案。
[1]
[1]
Another opposing perspective emphasizes the costs of fragmented regulation.
另一种相反的观点则强调分散监管的成本。
Reuters reports that state requirements and the federal administration’s deregulatory, innovation-focused approach are increasingly in conflict; the resulting patchwork may create uncertainty for companies and uneven protections for the public.
路透社报道称，州政府的要求与联邦政府放松管制、注重创新的做法日益冲突；由此产生的拼凑可能会给公司带来不确定性，并给公众带来不平衡的保护。
[2]
[2]
Get thought-provoking perspectives with our weekly newsletter.
通过我们的每周时事通讯获取发人深省的观点。
By continuing, you agree to our Terms of Service , which include arbitration and a class action waiver.
继续即表示您同意我们的服务条款，其中包括仲裁和集体诉讼豁免。
You agree that we and our third-party vendors may collect and use your information, including through cookies, pixels and similar technologies, for the purposes set forth in our Privacy Policy such as personalizing your experience and ads.
您同意我们和我们的第三方供应商可以出于我们的隐私政策中规定的目的（例如个性化您的体验和广告）收集和使用您的信息，包括通过 Cookie、像素和类似技术。