联邦政府将 AI 批评者视为“外国代理人”

对 AI 和数据中心的焦虑情绪广泛存在，但特朗普政府不知为何认定，公众的这种担忧是由中国制造出来的。

本周，特朗普总统发表了一番鲜少受到关注的言论，司法部也发出警告；政府由此向 AI 的抗议者和反对者宣战，威胁称，若有人“推进外国势力的宣传或其他目标”，将面临“刑事责任”。

这听起来仿佛是当代“东京玫瑰”在为北京张目，正通过短波电台接收指令。但实际目标只是普通人——他们完全不知道，自己已经成了政府最新国家安全幻象的一部分。

上周，司法部告知“公民和非公民”：任何人在“任何公共活动”中推进外国势力的“目标”——甚至包括“公开示威”——都必须正式通知政府，否则可能遭到逮捕和起诉。

尽管他们没有点名任何具体抗议活动，但不难猜出他们指的是什么。两天前，特朗普总统开始密集发布社交媒体帖子，宣称反对 AI 是一种可追溯至中国的叛国阴谋论。

“针对 AI 和数据中心，正发生一场令人作呕的阴谋，而唯一对此感到高兴的就是中国，”特朗普于9月14日在帖子中写道。“阴谋论者、叛国分子、卖国贼和泄密者，当心！”

特朗普还发帖称：“那些说 AI 将毁灭世界、说数据中心对你所在社区有害的人，都是革命者，但他们是在为邪恶且错误的事业革命。”

在第三篇帖子中，他将对 AI 的怀疑归咎于某个未具名的外国。“AI/数据中心风波之所以发生，唯一的原因是美国在领先——而且遥遥领先于其他所有国家，”他表示。

最近一次是在9月10日，特朗普暗示将利用刑事司法系统追查 AI 领域的“不良”行为者。他在9月19日警告：“我们也将寻找那些不良行为，而我们可以利用现有的刑事和民事司法系统非常轻易地做到这一点。”

换言之，总统正在打国家安全牌。而且不只是特朗普，国会也正在加入其中。

6月，参议院情报委员会主席汤姆·科顿（Tom Cotton，共和党，阿肯色州）要求时任代理司法部长托德·布兰奇（Todd Blanche）调查“针对美国 AI 基础设施建设的外国影响行动”。信中继续写道：“令人担忧的报告显示，一个由中国共产党（CCP）主导的外国行为者网络，正试图操纵美国有关数据中心的政策与公众舆论。”

科顿列举的主要证据是内维尔·罗伊·辛厄姆（Neville Roy Singham）。这名常驻上海的美国科技巨头拥有一个由美国左翼非营利组织构成的网络；科顿声称，该网络多年来一直制作反对美国 AI 基础设施的内容，而中国政府是其“最终出资方”。

“我写信请求司法部（DOJ）调查针对美国 AI 基础设施建设的外国影响行动，”参议院情报委员会主席汤姆·科顿（Tom Cotton，共和党，阿拉斯加州）6月致函时任代理司法部长托德·布兰奇时写道。“令人担忧的报告显示，一个由中国共产党（CCP）主导的外国行为者网络，正试图操纵美国有关数据中心的政策与公众舆论。”

（本通讯的读者或许还记得我去年的报道：一名国土安全官员告诉我，辛厄姆正根据总统的国家安全备忘录 NSPM-7 接受审查；该备忘录的说明见此处。）

科顿的信接着抱怨称：“[反 AI] 网络中没有任何实体因《外国代理人登记法》（FARA）而被起诉。因此，我请求司法部就这些事项展开全面调查。”

“我们不能容许外国对手借这些恐惧进行勒索，并破坏我们的技术发展，”参议院情报委员会主席汤姆·科顿（Tom Cotton，共和党，阿拉斯加州）6月致函时任代理司法部长托德·布兰奇时写道。科顿的信接着抱怨称：“[反 AI] 网络中没有任何实体因《外国代理人登记法》（FARA）而被起诉。因此，我请求司法部就这些事项展开全面调查。”

在科顿发信前几天，由主席布雷特·格思里（Brett Guthrie，共和党，肯塔基州）领导的众议院能源与商务委员会高级共和党议员，也提出了几乎完全相同的担忧。他们要求联邦调查局局长卡什·帕特尔（Kash Patel）和白宫就其所称、旨在阻止美国数据中心建设的外国影响行动进行简报。

“本届政府必须极其严肃地对待任何破坏[赢得 AI 竞赛]的行为——尤其是来自外国对手的行为，”他们写道。

格思里在自己的声明中表示：“中国共产党支持的实体及其他外国对手，可能正试图影响有关美国数据中心基础设施的决策；这一事实让我们认识到，我们正处于一场多么严峻的斗争中。”

在特朗普表态之前，他身边的一批民间人士早已在推动同一套论调。

先说大卫·萨克斯（David Sacks）。这名风险投资家曾担任特朗普的 AI 沙皇，后辞职并出任总统科学技术顾问委员会联席主席。萨克斯一直将美国国内对 AI 的抵制描绘为送给北京的大礼。7月，一个中国模型登上编程排行榜榜首后，他抱怨称：“美国正在自缚手脚：政客和官僚正禁止建设新的数据中心……这就是你输掉 AI 竞赛的方式。”最近，萨克斯还在福克斯新闻警告不要采取会“直接把整场比赛拱手让给中国”的行动。

还有凯文·奥利里（Kevin O’Leary），这位《创智赢家》（Shark Tank）投资人正在犹他州筹建一座投资额达1000亿美元的数据中心。奥利里是特朗普的坚定支持者，过去两年一直在拉近与政府的关系，其中包括2025年1月在海湖庄园与当选总统会面，以及在那里游说特朗普政府官员、试图收购 TikTok 的一段时期。

今年春季，奥利里声称，有数亿美元来自中国的资金正被用来支付抗议者，以反对像他这样的数据中心。他点名的一个组织“更好的犹他联盟”（Alliance for a Better Utah）回击道：“这个数据中心唯一的外国利益方，是来自加拿大的凯文。”

事情变得更可笑了。

在6月25日的一篇 X 帖子中，奥利里承认，他“没有证据”表明自己点名的组织和人士接受了中国资助；福克斯新闻也播出了道歉内容。其中两个组织此后已以诽谤为由起诉奥利里和福克斯新闻。

哎呀！

此外还有数量多到无法一一列举的行业组织。一个典型例子是：亲 AI 组织“创新委员会”（Innovation Council）的一名发言人告诉《每日呼声报》（Daily Caller），反数据中心运动是“一场由阴暗的反美势力和外国政府人为制造的行动”。

这一切都建立在一个前提上：如果不是北京在幕后操纵，普通美国人就不会反对数据中心。但民调表明并非如此。

盖洛普在3月进行的一项民调发现，71%的美国人反对在其所在地区建设 AI 数据中心。这一比例高于反对在本地建设核电站的53%。近半数人——48%——表示强烈反对，而且这种反对跨越党派界线，其中包括63%的共和党人。这并非边缘立场，而是接近共识。

这也不是技术恐惧症。根据皮尤研究中心2月对逾5000名成年人的一项调查，如今近半数美国人使用 AI 聊天机器人。但40%的人认为，AI 会在未来20年损害社会，相比之下仅16%认为它会带来帮助。人们不信任的是它被推出的方式：63%的人称 AI 发展过快，71%的人称它会降低其个人信息的安全性，67%的人对联邦政府监管 AI 几乎没有或完全没有信心。皮尤去年的研究发现，61%的人希望对 AI 在自己生活中的使用方式拥有更多控制权，约六成受访者担心政府监管会过于宽松。

换言之，公众希望拥有发言权，也希望至少建立一些管控措施。

特朗普政府的回应是：将反对 AI 和数据中心视为反情报事务。这并非政府首次打出国家安全牌，以为 AI 发展扫清道路。

6月，司法部介入全国有色人种协进会（NAACP）针对埃隆·马斯克 xAI 的诉讼。该诉讼涉及 xAI 在田纳西州运营数十台未取得空气排放许可的燃气涡轮机，为其维护的数据中心供电。司法部请求联邦法官驳回此案，称 NAACP “威胁美国的国家、经济和能源安全”。

为说明这一点，政府提交了五角大楼首席数字与人工智能官员卡梅伦·斯坦利（Cameron Stanley）的宣誓声明。斯坦利向法庭宣誓称，Grok “是至关重要的国家安全事项”。

将异议者说成中国傀儡的讽刺之处在于，这种做法很可能适得其反。当然，这只是其中一个讽刺。另一个是，特朗普一边警告“阴谋论者”，一边指控存在一场庞大的中国阴谋，但其盟友几乎没有拿出任何证据。（不过凯文·奥利里在诽谤诉讼的证据开示阶段，将有充足时间去寻找证据。）

当那些不信任政府处理 AI 方式的人被告知，他们的不信任是可被判处监禁的叛国性外国宣传时，很难想象这会让他们更加信任政府。更容易想象的是，他们会更加不信任政府，并开始怀疑：政府究竟如此急切地不想让他们问什么问题。这就是打国家安全牌的代价，而华盛顿似乎从来不会想到这一点。

CCP recruits young, apparently

Anxiety over AI and data centers is widely held, but somehow the Trump administration has convinced itself the public concern was manufactured in China.

This week, in little-noticed remarks by President Trump and Justice Department warning, the administration declared war on protestors and opponents of AI, threatening “criminal liability” if they “further the propaganda or other goals of a foreign power.”

While that sounds like some modern day Tokyo Rose shilling for Beijing, with orders coming in over shortwave radio, the actual targets are just normal people who have no idea they are part of the government’s latest national security mirage.

Last week, the Justice Department instructed “citizens and noncitizens” that anyone furthering the “goals” of a foreign power in “any public activity” — including even “public demonstrations” — must formally notify the government to avoid arrest and prosecution.

Though they don’t identify any specific protests, it’s not hard to surmise who they’re talking about. Two days prior, President Trump began a spate of social posts declaring that opposition to AI is a traitorous, treasonous conspiracy theory tracing back to China.

“There is a SICK conspiracy going on against AI and Data Centers, and the only one that is happy about it is China,” Trump posted on September 14. “Conspiracy Theorists, Treasonists, Traitors, and Leakers, BEWARE!”

“The people that say AI is going to destroy the World, and that Data Centers are bad for your neighborhood,” Trump also posted, “are Revolutionaries, but Revolutionaries for a Bad and Evil Cause.”

In a third post, he blamed the skepticism of AI on an unnamed foreign country. “The only reason the AI/Data Center outburst is happening is because the United States is leading, by a lot, every other country,” he said

Most recently, on September 10, Trump alluded to using the criminal justice system to go after “BAD” actors in the AI space. “[W]e will also be looking for BAD, and we can do that, very easily, with our already existing Criminal and Civil Justice System,” Trump warned on September 19.

In other words, the president is playing the national security card. And it’s not just Trump. Congress, too, is getting in on the act.

In June, Senate Intelligence Committee chairman Tom Cotton (R-AR) asked then-acting Attorney General Todd Blanche to investigate “foreign influence efforts targeting the buildout of American AI infrastructure.” The letter continues: “Alarming reports indicate that a network of foreign actors, led by the Chinese Communist Party (CCP), is attempting to manipulate U.S. policy and public opinion on data centers.”

Cotton’s chief exhibit was Neville Roy Singham, a Shanghai-based American tech mogul whose network of left-wing U.S. nonprofits, Cotton claimed, has spent years producing content opposing American AI infrastructure, with the Chinese government as its “ultimate paymaster.”

“I write requesting the Department of Justice (DOJ) investigate foreign influence efforts targeting the buildout of American AI infrastructure,” Senate Intelligence Committee chairman Tom Cotton (R-AK) wrote to then-acting attorney general Todd Blanche in June. “Alarming reports indicate that a network of foreign actors, led by the Chinese Communist Party (CCP), is attempting to manipulate U.S. policy and public opinion on data centers.”

(Readers of this newsletter will recall my report last year in which a homeland security official told me that Singham was being scrutinized under the president’s national security order NSPM-7, explained here.)

Cotton’s letter went on to complain that “no entity in the [anti-AI] network has been charged under the Foreign Agents Registration Act (FARA). I therefore request that the DOJ launch a full investigation into these matters.”

“We can’t allow any effort by foreign adversaries to extort these fears and undermine our technological development,” Senate Intelligence Committee chairman Tom Cotton (R-AK) wrote to then-acting attorney general Todd Blanche in June. Cotton’s letter went on to complain that “no entity in the [anti-AI] network has been charged under the Foreign Agents Registration Act (FARA). I therefore request that the DOJ launch a full investigation into these matters.”

Days before Cotton's letter, top Republicans on the House Energy and Commerce Committee, led by Chairman Brett Guthrie (R-KY), raised nearly identical concerns. They asked FBI Director Kash Patel and the White House for a briefing on what they called foreign influence campaigns aimed at blocking American data centers.

“It is critical that this Administration takes any effort to undermine [winning the AI race] — particularly from foreign adversaries — with a great deal of seriousness,” they wrote.

Guthrie said in his own statement: “The fact that Chinese Communist Party-backed entities and other foreign adversaries may be attempting to influence decisions related to American data center infrastructure puts into perspective how serious of a fight we are in.”

And well before Trump weighed in, a chorus of private figures in his orbit were pushing the same line.

Start with David Sacks, the venture capitalist who served as Trump’s AI czar before stepping down to become co-chair of the President's Council of Advisors on Science and Technology. Sacks has been casting domestic resistance to AI as a gift to Beijing. After a Chinese model topped a coding leaderboard in July, he complained that “America is tying itself in knots: politicians and bureaucrats are banning new data centers … This is how you lose the AI race.” More recently, Sacks warned on Fox News against moves that would “just hand the whole ball game to China.”

Then there’s Kevin O’Leary, the “Shark Tank” investor behind a planned $100 billion data center in Utah. O’Leary, a vocal Trump supporter, has spent the past two years cozying up to the administration. That includes a January 2025 Mar-a-Lago sit-down with the president-elect, and a stretch lobbying Trump officials there in a bid to buy TikTok.

This Spring, O’Leary claimed that hundreds of millions of dollars from China were paying protesters to oppose data centers like his. One group he targeted, the Alliance for a Better Utah, shot back: “The only foreign interest in this data center is Kevin from Canada.”

It gets funnier.

In a June 25 post on X, O’Leary conceded he had “no evidence” that the groups and people he’d named were funded by China, and Fox News aired apologies. Two of the groups have since sued O’Leary and Fox for defamation.

Whoops!

Then there are the industry groups, too numerous to list. A representative example: a spokesperson for the pro-AI Innovation Council told the Daily Caller that the anti-data center campaign was "a manufactured op by dark anti-American forces and foreign governments."

All of this rests on the premise that ordinary Americans wouldn’t oppose data centers unless someone in Beijing was pulling the strings. The polling says otherwise.

A Gallup poll conducted in March found that 71 percent of Americans oppose building an AI data center in their area. That’s more than the 53 percent who oppose a local nuclear plant. Nearly half, 48 percent, are strongly opposed, and the opposition crosses party lines, including 63 percent of Republicans. That’s not a fringe. It’s close to a consensus.

Nor is it technophobia. Nearly half of Americans now use AI chatbots, according to a February Pew surveyof more than 5,000 adults. Yet 40 percent said AI will hurt society over the next 20 years, compared with 16 percent who said it will help. What people distrust is how it’s being rolled out: 63 percent said AI is advancing too quickly, 71 percent said it will make their personal information less secure, and 67 percent had little or no confidence in the federal government to regulate it. Pew research last year found that 61 percent want more control over how AI is used in their lives, and about six in ten worry government regulation will be too lax.

In other words, the public wants a say, and it wants at least some controls in place.

The Trump administration’s response: treat opposition to AI and data centers as a counterintelligence matter. It’s not the first time the administration has played the national security card to clear the way for AI.

In June, the Justice Department intervened in an NAACP lawsuit against Elon Musk’s xAI over dozens of gas turbines the company was running without air permits to power a data center it maintained in Tennessee. The department asked a federal judge to throw the case out, arguing that the NAACP “threatens American national, economic, and energy security.”

To make the point, the government filed a sworn declaration from Cameron Stanley, the Pentagon’s chief digital and artificial intelligence officer. Stanley swore to the court that Grok was “a matter of paramount national security.”

The irony of the China puppetry gambit is that the approach is likely to backfire. Well, one of the ironies. Another is Trump warning about “Conspiracy Theorists” while alleging a vast Chinese plot for which his allies have produced virtually no evidence. (Though Kevin O'Leary will have plenty of time to find some during the discovery portion of his defamation suit.)

When people who distrust the government’s handling of AI are told their distrust is traitorous foreign propaganda punishable by jail time, it’s hard to imagine them trusting the government more. It’s easy to imagine them trusting it less, and wondering what the government is so eager to keep them from asking. That’s the downside of playing national security card, and it’s one that Washington never seems to think about.

Subscribe to make sure you see my next article, revealing how all of this will be formalized under a new government entity modeled off the military.

Leave a comment

Share

— Edited by William M. Arkin