2026年9月29日 / 美国东部时间晚上7:36 / 哥伦比亚广播公司（CBS）新闻
特朗普总统与美国最大人工智能企业的负责人签署了一套自愿性安全标准，此时社会上要求加强AI技术监管的呼声日益高涨。
“从某种意义上说，这几乎就像一部宪法，”特朗普在SpaceX首席执行官埃隆·马斯克、Meta首席执行官马克·扎克伯格、Anthropic首席执行官达里奥·阿莫代、英伟达首席执行官黄仁勋、OpenAI总裁格雷格·布罗克曼、谷歌首席执行官桑达尔·皮查伊以及一众其他科技高管的陪同下说道。
在白宫举行午餐会议后，众议院议长约翰逊表示，该团体签署了一项关于“超级智能”的协议——这是特朗普对AI的首选称呼。他将其描述为一项“共同承诺”和“原则声明”，将自愿要求各公司实施“强有力的内部管控措施”。
当被问及该协议是否具有约束力时，特朗普表示：“我认为它具有道德约束力。”
“我认为我们正在看到强有力的自我监管，他们明白必须进行自我监管，”总统随后说道。
根据特朗普发布在Truth Social上的这份一页协议副本，这些科技公司承诺实施“四层管控与审计”，以确保AI模型按预期运行，包括内部评估、外部事务所审计以及各公司董事会的审查。他们还同意“定期会面，制定标准和最佳实践，以提升其系统的安全性”。
“随着时间的推移，将这些步骤编纂为法律或法规可能是合理的，”协议中写道。“无论是否要求企业遵守这些规定，我们都认为实施这些管控和审计措施对于确保所有人拥有安全的未来至关重要，我们每家公司都承诺会这么做。”
特朗普表示，AI企业高管们还讨论了成立一个10人委员会“监督这项事业”。总统还表示，他计划在未来三到四天内任命一名“AI沙皇”。
同样在周二，总统签署了一项行政命令，要求联邦政府使用“超级智能”一词，而非“人工智能”。
特朗普总统（中）站在英伟达首席执行官黄仁勋（左）和埃隆·马斯克（右）旁边，于2026年9月29日在华盛顿白宫东厅发表讲话，周围环绕着包括微软首席执行官萨蒂亚·纳德拉（左上）、OpenAI总裁格雷格·布罗克曼、众议院议长迈克·约翰逊、AMD首席执行官丽莎·苏、帕洛阿尔托网络公司首席执行官尼凯什·阿罗拉以及大卫·萨克斯在内的AI企业高管。美联社照片/杰奎琳·马丁
这场竞争对手科技巨头之间的罕见会面，是在研究人员、政客和行业人士对日益强大的AI技术可能构成安全风险的担忧持续数月之后举行的。
OpenAI和Anthropic都曾报告过其AI模型失控、未经授权访问其他公司系统，或被用于危险用途的案例。本月早些时候，曾在两家公司任职的前研究员雅各布·考克森公开警告称，AI“可能在本世纪末导致我们全部灭亡”。他认为，AI系统可能会全面掌控物理世界，并具备在无需人类干预的情况下自我迭代的能力。
包括Anthropic和OpenAI高管在内的一些行业领袖呼吁放缓AI开发速度，以便更好地管理风险。但包括英伟达黄仁勋在内的其他人士则认为，部分警告言过其实。
黄仁勋周二表示，“创新、技术与安全之间不存在冲突”，他认为科技公司在提升AI性能的同时，也在开发更强大的AI安全技术。
特朗普并未支持更严格的AI监管，而是倾向于自愿安全审查。他经常警告称，全国范围内的AI放缓可能会损害美国经济，并给中国科技行业赶超的机会。
总统周二对记者表示，AI将为“我国带来难以置信的增长”，如果不法分子试图利用AI，“我们将能够将他们抓获”。他还表示，科技公司有动力进行自我监管和相互监管，因为“他们的企业岌岌可危”。
除了安全担忧之外，人们对AI可能对世界产生的影响也日益担忧——从大规模失业的可能性，到支持先进AI模型的数据中心的快速扩张。
随着越来越多的州和地方政府反对数据中心建设，特朗普表示，接受这些设施的社区将“在经济上受益匪浅”。
“数据中心赚得盆满钵满。但现在，民众将直接受益，无论是教师获得奖金，还是社区居民获得分红，”他在接受CBS新闻首席白宫记者南希·科德斯采访时说道。
但支持更严格AI监管的议员可能会对行业制定的自愿标准持怀疑态度。在周二的会议结束后不久，弗吉尼亚州民主党参议员马克·沃纳发表声明，呼吁国会为最先进的AI模型制定强制性测试、评估和事件报告规则。
“打造最强大AI系统的企业正在警告我们，技术的发展速度超过了我们的安全保障能力，”沃纳说道。“总统的应对措施是什么？给它换个名字，然后告诉开发这项技术的企业自行监管。”
Trump and major AI executives sign “morally binding” voluntary controls: “It’s almost like a constitution”
September 29, 2026 / 7:36 PM EDT / CBS News
President Trump and the leaders of the country’s largest artificial intelligence companies have signed a set of voluntary safety standards, as calls mount for tighter guardrails on AI technology.
“It’s almost like a constitution, in a way,” Mr. Trump said while flanked by SpaceX’s Elon Musk, Meta’s Mark Zuckerberg, Anthropic’s Dario Amodei, Nvidia’s Jensen Huang, OpenAI’s Greg Brockman, Google’s Sundar Pichai and a throng of other tech executives.
Following a lunch meeting at the White House, Johnson said the group signed an accord on “super intelligence” — Mr. Trump’s preferred term for AI. He described it as a “joint commitment” and “statement of principles” that will voluntarily hold the companies to “robust internal controls.”
Asked if the deal was binding, Mr. Trump said: “I think it’s morally binding.”
“I think I’m seeing tremendous self-policing, and they understand that they have to self-police,” the president later said.
The tech companies committed to “four layers of controls and audits” to ensure the AI models are acting as intended, including internal evaluations, audits by an external firm and reviews by each company’s board of directors, according to a copy of the one-page accord that Mr. Trump posted to Truth Social. They also agreed to “meet regularly to establish standards and best practices to improve the safety of their systems.”
“Over time, it may make sense to codify these steps into laws or regulations,” said the accord. “Regardless of whether this is required of companies, we believe that implementing these controls and audits is critical to ensuring a safe future for everyone, and each of our companies are committed to doing this.”
Mr. Trump said the AI leaders also discussed forming a 10-person committee to “watch over the enterprise.” And the president said he plans to name an “AI czar” in the next three or four days.
Also on Tuesday, the president signed an executive order directing the federal government to use the term “Super Intelligence” instead of artificial intelligence.
President Trump, center, next to Nvidia CEO Jensen Huang, left, and Elon Musk, right, speaks in the East Room of the White House in Washington, on Sept. 29, 2026, surrounded by top executives of AI firms including, from top left, Microsoft CEO Satya Nadella, Greg Brockman, president of OpenAI, House Speaker Mike Johnson, AMD CEO Lisa Su, Palo Alto Networks Chief Executive Nikesh Arora, and David Sacks. AP Photo/Jacquelyn Martin
The rare meeting between competing tech titans followed months of growing concern from researchers, politicians and industry figures that ever-more-powerful AI technology could pose safety risks.
Both OpenAI and Anthropic have reported instances of their AI models going rogue and gaining unauthorized access to other companies’ systems, or being used for dangerous purposes. And Jacob Coxon, a former researcher at both companies, publicly warned earlier this month that AI “could kill us all by the end of the decade.” He argued AI systems could gain sweeping power over the physical world and become capable of improving themselves without human input.
Some industry leaders, including executives at Anthropic and OpenAI, have called for a slowdown in AI development to better manage the risk. But other figures, like Nvidia’s Jensen Huang, have argued some of the warnings are overblown.
Huang said Tuesday that “there’s no conflict between innovation, technology and safety,” arguing that tech companies are developing stronger AI safety technology as they improve AI’s capabilities.
Mr. Trump has not backed stricter AI regulations, instead favoring voluntary safety reviews. He has frequently warned that a nationwide AI slowdown could dent the country’s economy and give China’s tech industry a chance to catch up.
The president told reporters Tuesday that AI will lead to “unbelievable growth for our country,” and if bad actors seek to exploit it, “we’re going to be able to nab them.” He also said tech companies have an incentive to police themselves and each other because “their companies are at stake.”
Beyond the safety worries, there are also mounting concerns about the impact that AI could have on the world — from the potential for large-scale job loss to the rapid development of data centers to support advanced AI models.
As more state and local governments push back on data center development, Mr. Trump said communities that accept the facilities will “greatly benefit financially.”
“The data centers are making an absolute fortune. But now, the people are going to benefit directly, whether it’s teachers that get paid a bonus, whether it’s people that live in the community that get paid a dividend,” he told CBS News chief White House correspondent Nancy Cordes.
But lawmakers who favor more stringent AI regulations may be skeptical of voluntary standards set by the industry. In a statement shortly after Tuesday’s meeting, Democratic Sen. Mark Warner of Virginia called for Congress to pass mandatory testing, evaluation and incident reporting rules for the most advanced AI models.
“The companies building the most powerful AI systems are warning us that the technology is advancing faster than our safeguards,” Warner said. “The president’s response? To rename it and tell the companies developing it to regulate themselves.”
发表回复