在华盛顿，推动对人工智能（AI）进行监管的努力遇到了重大障碍——那些主张严格执行反垄断法规的议员们对此表示强烈反对。
The push to regulate artificial intelligence faces a big speed bump in Washington Antitrust hawks.
AI企业要求获得有限的反垄断豁免权，以便在安全问题上展开合作。随着人们对AI技术可能失控的担忧日益加剧，这些企业不断呼吁政府加强对AI行业的监管。然而，负责制定相关法规的国会议员们对这种豁免请求持反对态度。
AI companies have asked for narrow antitrust exemptions to collaborate on safety issues as they push Washington to regulate them amid growing fears that the technology could grow out of control But some lawmakers on Capitol Hill who would write the laws regulating the industry are recoiling at the suggestion of an antitrust carveout. "No," Sen. Elizabeth Warren D-Mass., told CNBC at the Capitol when asked about a potential waiver.
当被问及是否可能给予此类豁免时，马萨诸塞州民主党参议员伊丽莎白·沃伦（Elizabeth Warren）对CNBC表示：“反垄断法已经为那些在安全问题上合作的企业提供了相应的豁免条款，因此没有必要再额外授予任何豁免权。”她补充说：“我知道大企业很希望获得反垄断豁免，但我的答案是否定的。”
"Antitrust law already has an exception for companies working together on safety issues; there's no need for any additional antitrust waivers." "I understand that big companies love antitrust waivers, but no," she said. Warren was one of a number of antitrust exemption skeptics on Capitol Hill who CNBC spoke with to gauge the appetite for granting the request floated by Anthropic Dario Amodei with support from other AI executives.
沃伦是国会中众多反对给予此类豁免的议员之一。CNBC还采访了其他几位议员，以了解他们对这一请求的态度。
In a blog post earlier this month, Amodei raised the possibility of an antitrust waiver.
本月早些时候，Anthropic公司的首席执行官达里奥·阿莫迪（Dario Amodei）在一篇题为《我们必须把握技术发展的节奏》（“We Must Pace the Frontier”）的博客文章中提出了这一请求。这篇文章的背景是：Anthropic的一名研究人员辞职，并公开表示AI技术“可能在本十年内导致人类灭亡”。
The essay, titled "We Must Pace the Frontier," came after one of Anthropic's researchers resigned and said publicly that AI "could kill us all by the end of the decade." Trump holds up to $1 billion in bonds tied to institutions he can affect Rep. James Comer expands House investigation into prediction market insider trading Hassett: 'We're still working our way out from' Biden Covid spending "AI companies can and should work together to set standards," Amodei wrote in the essay.
特朗普持有价值高达10亿美元的债券，这些债券与某些机构有关；众议员詹姆斯·科默（James Comer）因此要求扩大对AI预测市场内幕交易的调查范围。阿莫迪在文章中写道：“AI企业完全可以也应该共同努力来制定行业标准。出于反垄断考虑，美国政府应该促成或至少支持这些讨论——虽然政府无需直接参与其中，但至少应为某些与安全相关的讨论提供必要的豁免权限。”其他AI企业的高管也迅速支持了阿莫迪的提议，这促使华盛顿方面加快了制定AI行业监管规则的进程。
"For antitrust reasons, it's helpful for the US government to mediate or at least enable these discussions — they don't need to participate, but do need to issue a narrow waiver for certain kinds of safety conversations." Other AI executives quickly endorsed Amodei's call to action, setting off a mad dash in Washington to come up with rules for the technology.
但今年留给国会立法的时间所剩无几，加上政府立法部门内部严重分裂，任何正在酝酿的不满情绪，都可能威胁到围绕人工智能这一关键议题迅速通过立法。
But the short amount of time left this year for Congress to legislate and the bitterly divided nature of the government's legislative branch mean that any brewing dissent presents a risk to quickly passing legislation on a pivotal topic like AI.
反垄断执法机构同样怀疑是否应给予人工智能行业任何例外安排；该行业的高管周二曾在白宫与特朗普会面。
Antitrust enforcers are also skeptical of any carveouts for AI, whose executives met Tuesday at the White House with President Donald Trump.
除阿莫代外，谷歌首席执行官桑达尔·皮查伊、特斯拉和SpaceX首席执行官埃隆·马斯克、Meta首席执行官马克·扎克伯格、OpenAI总裁格雷格·布罗克曼，以及英伟达的黄仁勋也出现在白宫；特朗普则宣布，人工智能应改称为“超级智能”。
In addition to Amodei Google Sundar Pichai Tesla SpaceX chief Elon Musk Meta Mark Zuckerberg OpenAI President Greg Brockman and Jensen Huang of Nvidia appeared at the White House as Trump declared that artificial intelligence should be called "super intelligence" instead.
曾在前总统乔·拜登任内负责司法部反垄断部门的乔纳森·坎特告诉CNBC，让这些公司为安全事项开展合作，无须豁免现行反垄断法。
Jonathan Kanter who ran the Justice Department's antitrust division under former President Joe Biden, told CNBC an exemption to existing antitrust laws is not needed for the companies to collaborate on safety.
“它们可以共享有关恶意威胁的信息以及其他信息，而无须获得反垄断法豁免，”坎特说。“反垄断法并不阻止公司聚在一起讨论如何制定标准，让产品更加安全……反垄断法可能阻止的是公司达成共识，认为它们应该停止创新、停止竞争。”
"They can share information about malicious threats and other information without an antitrust exemption," Kanter said. "The antitrust laws don't prevent companies from coming together to talk about how to develop standards to make products more safe ... what the antitrust laws might prevent is companies coming together and saying they should stop innovating and stop competing."
阿莫代在其讨论人工智能发展节奏的文章中，将这种放缓描述为集体“对不受节制的人工智能进展速度设定限制”，并指出这样做可能带来法律上的挑战。“它们不需要联合作出这些决定，”坎特说。
Amodei described such a slowdown in his essay about pacing as collective "limits on the rate of unchecked AI progress" and noted they could be legally challenging. "They don't need to make those decisions jointly," Kanter said.
坎特表示，国会应另行考虑制定“一些基本规则，确保公众在这个新兴领域免受有害产品和服务的侵害”。他还警告称，安全可能成为 companies 协调实施其他类型反竞争行为的借口。
Kanter said Congress should separately consider imposing "basic rules of the road to ensure that the public is protected from harmful products and services in this new and emerging area." He also warned safety could be used as a pretext for the companies to coordinate on other types of anticompetitive behavior.
“把反垄断法豁免用错，确实存在很大风险，”他说。“话虽如此，我认为并不需要这种豁免。
"There's a real risk that you can get the antitrust exemption wrong," he said. "That being said, I don't think you need one.
坎特（Kanter）和沃伦（Warren）并不是唯一认为这些公司在进行安全协调时并不需要反垄断豁免的人。
Kanter and Warren are not alone in saying they don't think antitrust waivers are necessary for the companies to coordinate on safety.
“只要它们遵守政府的某些规定或接受监管，就不需要反垄断豁免来避免灾难性的后果，”康涅狄格州民主党参议员理查德·布卢门撒尔（Richard Blumenthal）对CNBC表示。
"They don't need a waiver to prevent catastrophic consequences as long as they're cooperating with some kind of governmental rubric or oversight," Sen. Richard Blumenthal D-Conn., told CNBC.
亚利桑那州民主党参议员鲁本·加列戈（Ruben Gallego）则警告说，国会可能会草率地给予人工智能行业一切它想要的东西。
And Sen. Ruben Gallego D-Ariz. warned about the possibility that Congress could rush into delivering the AI industry everything it wants.
“首先，这样的立法并不能免除这些公司的法律责任，”加列戈说。“我不会允许它们在所有事情上都拥有完全的自由裁量权。”
"I'd have to see exactly how far that goes, it doesn't remove them from liability issues, first of all," Gallego said of potential legislation.
北达科他州共和党参议员约翰·霍文（John Hoeven）表示，国会仍在努力寻找一种协调一致的方式来应对人工智能领域的发展。他警告称，过去在监管互联网等新兴技术时的经验对立法者来说具有重要的参考价值。
"I'm not going to give them a free carte blanche on everything." John Hoeven, R-N.D., said Congress is still trying to figure out a coordinated way forward on addressing AI. He warned that prior experience regulating other recent technological breakthroughs like the internet looms large over lawmakers.
“对于互联网企业所获得的那些反垄断豁免条款，人们普遍感到不满，”霍文对CNBC说，“因此我们正在思考如何避免类似的问题在人工智能领域再次发生。单纯给予它们反垄断豁免权是行不通的。”
"I think there's unhappiness with these antitrust carveouts that the internet guys got, so we're thinking about, okay, how do we avoid that problem with AI?" Hoeven told CNBC. "To just give them some carte blanche antitrust thing, we've seen that doesn't work."
除了反垄断方面的担忧外，一些立法者还对要求放缓人工智能发展的呼声持怀疑态度，担心这可能导致监管机构被企业势力所控制（即监管机构本应服务于公众利益，却反而屈从于企业的利益）。虽然目前还没有确凿的证据表明存在这种协调行为，但一些分析师认为，这种做法可能是企业为在首次公开募股前控制自身不断增长的债务负担、从而抑制市场竞争而采取的手段。
In addition to the antitrust concern, some lawmakers are suspicious of calls for a coordinated slowdown of AI, warning it could result in regulatory capture — where a regulator intended to serve the public's needs instead bends to industry. Some analysts have suggested it could be a ruse to curb the companies' growing debt portfolios ahead of initial public offerings and stifle competition though there is not concrete evidence of such coordination.
“有一点吧，”北达科他州共和党参议员凯文·克莱默周二在CNBC的《Squawk Box》节目中被问及是否担心监管机构可能被某些大型企业所“俘获”（即被企业影响而偏离公正的监管方向）时回答道。“任何阻碍大型企业进入某个行业的行为，我都会感到怀疑。”虽然人工智能行业在国会中得到了一些支持，但该行业仍呼吁获得特定的反垄断豁免权。
"A little bit," Sen. Kevin Cramer, R-N.D., said Tuesday when asked about whether he's concerned about the potential for regulatory capture on CNBC's "Squawk Box." "Anything that blocks entrance to an industry by somebody who's big in it, again, I'm always a little bit suspicious of." AI companies have some congressional support Industry's call for a narrow antitrust exemption does have some supporters. Reps.
加州民主党众议员乔治·怀特赛德斯和俄亥俄州共和党众议员鲍勃·拉塔，以及加州民主党参议员亚当·希夫和印第安纳州共和党参议员吉姆·班克斯，共同提出了一项法案，旨在为人工智能企业提供这种豁免权。该法案允许这些企业共享与人工智能安全威胁相关的信息，以便共同预防、调查或减轻相关风险；同时，这些企业还可以通过推迟或限制高风险人工智能模型的部署来协作解决安全问题。
George Whitesides D-Calif. and Bob Latta, R-Ohio, and Sens. Adam Schiff D-Calif., and Jim Banks, R-Ind., carry a bill that would create such a carveout. It would allow AI companies to share information relating to AI security threats to prevent, investigate or mitigate the risk. The bill would also allow them to coordinate to reduce AI security issues by delaying or limiting the deployment of high-risk models.
“这种豁免权必须被严格限制在使用范围，以确保其仅用于那些与人工智能模型的测试、评估和训练直接相关的工作，从而提高模型的安全性，”推动这项法案的加州民主党众议员萨姆·利卡多对CNBC表示。“坦率地说，这些企业几乎没有动力去进行广泛的合作——毕竟这是一个竞争极其激烈的行业。”请将CNBC设为您在谷歌上的首选新闻来源，这样您就不会错过任何一条来自这一权威商业新闻媒体的重要消息。
"It needs to be harbored very narrowly to ensure the focus is on exactly what needs to be done around testing, evaluation and training that makes these models safer," Rep. Sam Liccardo D-Calif., who pushed for Whitesides' bill, told CNBC. "Frankly, there's not a lot of incentive for these companies to want to collaborate on a whole lot of things; this is an intensely competitive industry." Choose CNBC as your preferred source on Google and never miss a moment from the most trusted name in business news.