How does Trump’s White House AI accord work?

Trump announces voluntary pact on AI, urging companies to self-police amid growing calls for strict safety regulations.

Executives from Meta, Nvidia, xAI, OpenAI, Google and Anthropic met with United States President Donald Trump on Tuesday [Jonathan Ernst/Reuters]

Amid calls for regulation, United States President Donald Trump met with artificial intelligence executives at the White House to put in place a voluntary pact, called the White House Accord on Super Intelligence.

The agreement, which was signed by executives at Meta, Nvidia, Google, OpenAI, xAI and Anthropic on Tuesday, calls on companies to put in place guardrails internally to help monitor security risks, with four layers of control.

list 1 of 4 Trump backs AI self-regulation at tech summit but is it enough?

list 2 of 4 OpenAI launches ‘dots,’ personal AI assistant ‘built to handle everything’

list 3 of 4 Trump unveils new site to simplify access to government services

list 4 of 4 Trump, tech bosses sign voluntary pact pledging ‘robust’ AI safeguards

What does the White House Accord do and what doesn’t it do?

The agreement calls for companies to maintain internal controls, establish an internal team to oversee these controls, work with independent outside auditors and create an independent board committee to review the results.

The pact, however, is voluntary and not legally binding. It does not establish penalties for companies that fail to comply. The pact does not require companies to publicly disclose audit results or give the government an enforcement role. Companies also retain control over how the safeguards are implemented.

The agreement does, however, leave open the possibility that its measures could eventually become mandatory. It says that, over time, it might “make sense to codify these steps into laws or regulations”.

“There’s going to be a tremendous self-policing aspect,” Trump told reporters, adding that it is “morally binding”.

On the White House lawn, alongside the executives, Trump claimed that AI companies had done a great job at self-policing.

Google CEO Sundar Pichai also touted the agreement on social media.

“We are committed to working with other industry leaders to establish norms and build public confidence. The White House Accord and the Joint Commitment on Frontier Responsibilities signed today is a solid basis for moving forward – it contains real tangible steps to promote safe development, while delivering the economic and scientific benefits of this technology,” Pichai said in a post on X.

Why are AI companies calling for more regulation?

The agreement comes despite calls by some of the same companies for government-mandated safety standards.

In September, Chris Lehane, chief global affairs officer at OpenAI, the maker of ChatGPT, called for Congress to establish mandatory national AI safety standards.

“Industry-led standards would complement – not replace – mandatory federal safeguards and democratic oversight,” Lehane said in a post.

In June, Anthropic CEO Dario Amodei called for “more serious and binding regulation of AI”. In July, Demis Hassabis, CEO of Google DeepMind, called for an AI watchdog led by the US.

The calls for guardrails come amid reports that some employees at those companies have raised concerns about whether increasingly sophisticated AI systems are being adequately watched.

Two employees at OpenAI warned executives that testing of the company’s new models was not being properly monitored. The warnings were reportedly not acted upon ahead of incidents in which the models behaved in unintended ways, according to The New York Times.

What are lawmakers proposing?

Pushes for legally binding regulations are coming from Congress rather than the White House.

Earlier this month, Jacob Coxon, a former AI researcher at Anthropic, warned that AI technology could pose an existential threat to humanity within the next decade, prompting renewed attention from lawmakers.

Among the most comprehensive proposals comes from Independent Vermont Senator Bernie Sanders and Representative Greg Casar, a Democrat from Texas.

The two lawmakers introduced the Ban Artificial Superintelligence Act, which, if passed, would permanently ban the development and deployment of artificial superintelligence and temporarily pause advanced AI development until a federal regulator establishes safety rules.

The bill would also create a cabinet-level Department of Artificial Intelligence to monitor frontier AI systems – a term used to describe the most advanced AI systems – as well as oversee dangerous capabilities and supervise the destruction of artificial superintelligence if needed.

The legislation would impose harsh penalties for violations, such as up to 20 years in prison for individuals.

It comes amid a bipartisan bill introduced last week by Senators Chris Coons, a Democrat from Delaware; Brian Schatz, a Democrat from Hawaii; Katie Britt, a Republican from Alabama; and James Lankford, a Republican from Oklahoma.

The AI Systems Transparency Act would require AI companies to disclose what data they collect and what measures they have in place to prevent their systems from going rogue or being misused by bad actors.

What does ‘superintelligence’ actually mean?

The White House named the accord the White House Accord on Super Intelligence amid Trump’s push to rename “artificial intelligence” on government documents.

In the technology industry, “superintelligence” refers to a form of AI that exceeds human intellectual capabilities.

Trump, however, has directed executive branch agencies to use “Super Intelligence” or “SI” in place of “artificial intelligence” or “AI” in official communications.

On Tuesday, Trump outlined a new executive order that says that “Super Intelligence” refers to technologies and systems covered by the existing statutory definition of artificial intelligence and directs the creation of a new federal definition within 60 days.

特朗普宣布就人工智能达成自愿协议，敦促各公司进行自我监管，同时要求制定严格安全法规的呼声日益高涨。

Meta、Nvidia、xAI、OpenAI、Google 和 Anthropic 的高管周二与美国总统唐纳德·特朗普会面。[Jonathan Ernst/路透社]

在各方呼吁加强监管之际，美国总统唐纳德·特朗普在白宫会见了人工智能领域的高管，达成了一项名为《白宫超级智能协议》的自愿协议。

该协议由 Meta、Nvidia、Google、OpenAI、xAI 和 Anthropic 的高管于周二签署，要求各公司在内部建立防护措施，以帮助监控安全风险，并设有四层控制措施。

列表 1/4：特朗普在科技峰会上支持人工智能自我监管，但这足够吗？

OpenAI推出“dots”，一款“旨在处理一切”的个人AI助手（共4条列表，此为第2条）。

特朗普推出新网站，简化获取政府服务的流程（共 4 条，此为第 3 条）

列表 4/4：特朗普与科技公司高管签署自愿协议，承诺提供“强有力的”人工智能安全保障

《白宫协议》的作用是什么？它又没有起到什么作用？

该协议要求各公司维持内部控制，建立内部团队来监督这些控制措施，与独立的外部审计师合作，并设立独立的董事会委员会来审查结果。

然而，该协议是自愿性的，不具有法律约束力。它没有规定对不遵守协议的公司进行处罚，也没有要求公司公开披露审计结果，更没有赋予政府执法权。公司还可以自主决定如何实施这些保障措施。

然而，该协议并未排除其措施最终可能成为强制性措施的可能性。协议指出，随着时间的推移，“将这些步骤编纂成法律或法规或许是合理的”。

特朗普告诉记者：“将会有很强的自我监管机制”，并补充说这“具有道德约束力”。

在白宫草坪上，特朗普与几位高管一起声称，人工智能公司在自我监管方面做得非常出色。

谷歌首席执行官桑达尔·皮查伊也在社交媒体上大力宣传了这项协议。

“我们致力于与其他行业领导者合作，建立规范并建立公众信心。今天签署的《白宫协议》和《前沿责任联合承诺》为向前迈进奠定了坚实的基础——它包含了促进安全开发的切实可行的步骤，同时带来了这项技术的经济和科学效益，”皮查伊在X网站上发帖称。

为什么人工智能公司呼吁加强监管？

尽管其中一些公司呼吁政府强制制定安全标准，但双方还是达成了协议。

9 月，ChatGPT 的开发商 OpenAI 的首席全球事务官 Chris Lehane 呼吁国会制定强制性的国家人工智能安全标准。

莱恩在一篇帖子中表示：“行业主导的标准将作为补充，而不是取代强制性的联邦保障措施和民主监督。”

今年6月，Anthropic首席执行官达里奥·阿莫迪呼吁“对人工智能进行更严格、更具约束力的监管”。7月，谷歌DeepMind首席执行官德米斯·哈萨比斯呼吁设立由美国牵头的人工智能监管机构。

有报道称，这些公司的一些员工对日益复杂的人工智能系统是否得到充分监控表示担忧，与此同时，各界也呼吁设立防护措施。

据《纽约时报》报道，OpenAI的两名员工曾警告高管，公司新模型的测试没有得到妥善监控。然而，在模型出现意外行为之前，这些警告并未得到重视。

立法者们提出了什么方案？

推动制定具有法律约束力的法规的动力来自国会，而不是白宫。

本月初，人智公司前人工智能研究员雅各布·考克森警告说，人工智能技术可能在未来十年内对人类构成生存威胁，这引起了立法者的重新关注。

其中最全面的提案来自佛蒙特州独立参议员伯尼·桑德斯和德克萨斯州民主党众议员格雷格·卡萨尔。

这两位议员提出了《禁止人工智能超级智能法案》，如果该法案获得通过，将永久禁止开发和部署人工智能超级智能，并暂时中止高级人工智能的开发，直到联邦监管机构制定安全规则。

该法案还将设立一个内阁级别的人工智能部，负责监控前沿人工智能系统（该术语用于描述最先进的人工智能系统），以及监督危险能力，并在必要时监督销毁超级人工智能。

该法案将对违法行为处以严厉的处罚，例如对个人处以最高 20 年的监禁。

此前，特拉华州民主党参议员克里斯·库恩斯、夏威夷州民主党参议员布莱恩·沙茨、阿拉巴马州共和党参议员凯蒂·布里特和俄克拉荷马州共和党参议员詹姆斯·兰克福德上周提出了一项两党法案。

《人工智能系统透明度法案》将要求人工智能公司披露其收集的数据以及为防止其系统失控或被不良行为者滥用而采取的措施。

“超级智能”究竟是什么意思？

在特朗普推动政府文件中“人工智能”更名之际，白宫将该协议命名为《白宫超级智能协议》。

在科技行业中，“超级智能”指的是超越人类智力能力的人工智能形式。

然而，特朗普已指示行政部门机构在官方沟通中使用“超级智能”或“SI”来代替“人工智能”或“AI”。

周二，特朗普概述了一项新的行政命令，该命令称“超级智能”指的是现有人工智能法定定义所涵盖的技术和系统，并指示在 60 天内制定新的联邦定义。