参议员乔什·霍利（共和党，密苏里州）和克里斯·墨菲（民主党，康涅狄格州）计划共同提出一项关于人工智能责任的立法法案。在国会山，各方就如何规范这项技术仍存在分歧，但两党之间的合作努力仍在持续。
Sens. Josh Hawley (R-Mo.) and Chris Murphy (D-Conn.) are planning to introduce AI liability legislation asbipartisan efforts to regulate the technology persist on Capitol Hill. Why it matters: While President Trump has made his administration's position abundantly clear — the industry should self-regulate — a growing number of congressional Republicans insist new laws are needed on the books. TheAI Agent Accountability Act will seek to hold companies criminally and civilly liable for hacking incidents, Axios has learned.
为何这很重要：尽管特朗普总统已明确表示政府立场——即人工智能行业应自行进行自我监管——但越来越多的共和党议员坚持认为需要制定新的法律来规范这一领域。据Axios报道，《人工智能代理责任法案》旨在追究企业在人工智能系统被黑客攻击事件中的刑事和民事责任。
The lawmakers argue that existing law leaves too much uncertainty over who can be held liable when an AI agent hacks a system.
这些议员认为，现行法律在界定谁应承担责任方面存在太多不确定性。他们的立场与政府官员的观点形成鲜明对比；政府官员认为现有法律足以应对人工智能可能带来的危害。
The approach is in stark contrast from administration officials, who say existing laws can address AI harms. "As the president so noted, we have consumer protection laws. We have product liability laws," Director of National Intelligence Jay Clayton said Wednesday on Squawk Box.
“正如总统所说，我们已有消费者保护法律和产品责任法律，”美国国家情报总监杰伊·克莱顿周三在Squawk Box节目中表示，“我们还有司法部，以及处理交通、能源、金融服务等领域发展问题的相关机制。”特朗普认为克莱顿是担任人工智能监管机构的合适人选。
"We have the [Department of Justice]. We have architectures for dealing with developments in all of the verticals: transportation, energy, financial services, and the like," said Clayton, who Trump views as a good option for AI czar. The big picture: Lawmakers on both sides of the aisle aren't sold and continue to introduce bills. Reps.
整体形势：两党议员均对现有法律感到不满，因此继续推动相关立法进程。加州民主党议员泰德·刘（Ted Lieu）和德克萨斯州共和党议员纳撒尼尔·莫兰（Nathaniel Moran）提出了一项法案，要求开发者必须建立能够关闭具有灾难性风险的人工智能系统的机制；新泽西州共和党议员汤姆·基恩（Tom Kean）和路易斯安那州共和党参议员约翰·肯尼迪（John Kennedy）也提出了类似的法案。此外，还有许多其他涉及人工智能模型测试、风险管理、透明度及网络安全等方面的两党联合立法提案。
Ted Lieu (D-Calif.) and Nathaniel Moran (R-Texas) have a bill that would require developers to have a mechanism for shutting down AI that poses catastrophic risk. Rep. Tom Kean (R-N.J.) and Sen. John Kennedy (R-La.) also introduced similar kill switch legislation. Many other bipartisan AI bills have been introduced around model testing, risk management, transparency and cybersecurity.
关键在于国会领导层：值得注意的是，参议院多数党领袖约翰·图恩（John Thune，共和党，南达科他州）本周并未出席白宫举办的关于人工智能的午餐会。当时参议院仍在开会，图恩反而与Anthropic公司的首席执行官达里奥·阿莫迪（Dario Amodei）以及白宫幕僚长苏西·怀尔斯（Susie Wiles）进行了会面。
The key is Hill leadership. Senate Majority Leader John Thune (R-S.D.)was notably absent from a White House AI lunch this week. The Senate was still in session and he instead met with Anthropic CEO Dario Amodei and White House chief of staff Susie Wiles there.
关于人工智能相关立法的谈判陷入了僵局，因为国会正忙于为选举季做准备。
Negotiations on his AI legislation are stalled as Congress taps out for election season.
众议院议长迈克·约翰逊（共和党-路易斯安那州）在午餐后对记者表示，由于技术发展速度过快，国会此前所做的许多跨党派努力都已“变得过时”。他补充说，在业界及立法者之间达成共识非常困难。
House Speaker Mike Johnson (R-La.) told reporters following the lunch that much of Congress' bipartisan work has been "rendered obsolete" because of how fast the technology moves. He added it's difficult to build consensus in industry and among lawmakers.
不过，一家与立法工作相关的咨询公司代表表示：“法律规定的透明度措施能够跟上人工智能发展的步伐，因为这些措施并不依赖于技术的进步。”
Yes, but: Legally mandated transparency measures can keep pace with AI development because they operate independently of technological advances, a benchmarking company representative who recently met with lawmakers said.
一位熟悉该立法进程的消息人士指出：“霍利（Hawley）的团队认为，追究因人工智能技术造成的损害的责任要比试图制定一套适用于不断变化的技术的监管制度要简单得多。”
Hawley's office sees liability for harm as simpler than trying to write a regulatory regime on technology that is always changing, said a source familiar with the legislation.
归根结底：只要特朗普及其在国会的盟友继续听从业界的意见，这种缺乏共识的局面就必然会影响到国会决策过程。
The bottom line: So long as Trump and his Hill allies take their cues from industry, its lack of consensus will inevitably spill into Congress.