AI-powered robocalls attack Talarico on data centers

Kaylee Greenlee/Reuters

Lone Star Liberty, a pro-Ken Paxton super PAC that received $10 million from Elon Musk , is using an AI-powered robocall campaign in Texas to attack Democratic Senate candidate James Talarico for allegedly taking money from “data center special interest” groups, according to audio recordings reviewed by CNN.

The calls, which have not been previously reported, feature an AI-generated caller named “Jennifer” who identifies itself as a virtual staffer for Lone Star Liberty. Unlike a traditional pre-recorded robocall, the system is designed to interact with voters in real time, asking whom they supported and responding to their answers.

“Hi, is this Julio? Hi. My name is Jennifer and I’m a virtual staffer with Lone Star Liberty calling in support of Republican Ken Paxton,” the AI-generated voice says in one call reviewed by CNN.

"Hi, is this Julio? Hi. My name is Jennifer and I’m a virtual staffer"

Lone Star Liberty, a super PAC supporting Ken Paxton in his campaign to win a seat in the US Senate, is using an AI-powered robocall campaign in Texas to attack Democratic Senate nominee James Talarico.

The AI-voice delivers Republican attack lines against Talarico – including alleging in one call that he “took $11 million from data center special interests.”

AI robocaller attacks James Talarico on...data centers

The Texas super PAC is using an AI tool to attack Talarico for taking money from the very data centers that make those kinds of AI calls possible.

A spokesperson for the PAC said the figure was based on a Daily Caller report that detailed donations to pro-Talarico political groups from individuals with ties to the data center industry. Those donors included tech billionaire Reid Hoffman, who has championed data center expansion despite widespread public opposition , and who contributed $10 million to a super PAC supporting Talarico this cycle.

The robocall campaign presents a striking irony: Musk, the world’s richest person and the CEO of a top AI company, is spending hundreds of millions of his own fortune to support Republican political campaigns and super PACs. In this case, the Texas super PAC is using an AI tool to attack taking money from the very data centers that make those kinds of calls possible.

AI robocalls are not exclusive to political campaigns — other industries such as insurance and finance are adopting them — but the calls also offer an early look at how campaigns can use conversational AI calls to conduct political persuasion at a potentially massive scale — an emerging form of political outreach that is difficult for the public to track, in part because it lacks a centralized database.

CNN’s KFile identified the robocalls through Nomorobo, a robocall-and text-blocking application, and obtained recordings of the calls. The calls were first detected on October 23, originating from dozens of phone numbers and reaching a relatively small number of landlines, according to Nomorobo.

Data centers have become one of the most potent political targets of the 2026 election cycle, drawing opposition across party lines over their energy use, costs and rapid expansion. Recent CNN polling found that more than 8 in 10 Americans expressed far more concern than excitement about the rapid development of artificial intelligence.

Brandon Bell/Getty Images

Yet even as voters voice misgivings, political campaigns are embracing AI-generated political TV and digital ads, and an emerging frontier of AI powered conversational robocalls.

In a statement to CNN, Gregg Keller, a spokesperson for Lone Star Liberty, defended the group’s campaign tactics, saying it was “using every tool available” to inform Texas voters of their criticisms of Talarico.

A spokesperson for the Talarico campaign told CNN in a statement, “Ken Paxton can’t get humans to make calls for him, so Elon Musk is firing up his data centers to create the only volunteers willing to spread lies for their campaign: AI robots.”

Nate Persily, a professor at Stanford Law School and the co-director of the Stanford Law AI Initiative, told CNN in an email he is skeptical that AI robocalls impact political races since people rarely pick up their phones and might likely hang up if they hear a “virtual staffer.”

“Most people see robocalls as annoying – whether the traditional recorded ones or now, the AI-generated ones,” said Persily. “We ought to separate out false or polarizing claims made by a communication and the technology of that communication. I appreciate that AI is the shiny new object of this campaign but manipulation is as old as politics.”

The central question for AI-powered robocall technology is “whether it is cheaper and more effective at manipulation and whether it can do so at scale,” he added.

While these kinds of AI-powered robocalls are at the moment generally limited to people’s landlines, their use could become far more prevalent depending on how the Federal Communications Commission rules on a petition now under consideration.

Currently, federal law bans most robocalls to wireless phones unless the caller has consent beforehand. The FCC is considering whether to loosen its rules on political robocalls — a change sought by the conservative group Club for Growth that could permit large-scale unsolicited calls using artificial or prerecorded voices, to reach not just landlines but mobile phones without prior consent.

Top Democrats and consumer-rights groups oppose the petition. The FCC could rule by the end of October. The debate comes more than two years after the FCC declared robocalls using AI to clone someone’s voice as illegal under existing federal law after a call that impersonated President Joe Biden and urged New Hampshire voters not to cast ballots was used in the 2024 primary.

The recordings reviewed by CNN show both the promise and the limits of the technology. The AI caller can react to a recipient, but it also glitches when interrupted— stopping, restarting, and repeating questions or campaign talking points. Nomorobo removed recipients’ voices from the recordings reviewed by CNN in order to protect their privacy.

If asked who it supports in the race, the virtual staffer responds, “I don’t have personal opinions because I’m an AI virtual staffer. I can explain the campaign’s position, but the decision is yours.”

"I don’t have personal opinions because I’m an AI virtual staffer"

The AI robocaller engages with call recipients and can answer questions, but when asked about its opinions on the race, it declines.

A CNN review of robocalls with the term “virtual staffer” found that the same “Jennifer” AI voice was used by at least two other PACs this cycle.

The Missouri Promise PAC , established earlier this year and funded by dark money (funds from groups that do not disclose their donors), used AI robocalls this summer to ask about Amendment 5, which would give lawmakers authority to phase out Missouri’s state income tax and replace that revenue by increasing sales taxes. Voters ultimately rejected the measure in August.

The same AI robocall voice launched another campaign this October with “Voters for Integrity,” a Missouri PAC established in August in support of multiple Republican candidates running for seats in the state.

But the recordings from neither PAC’s AI robocalls mention data centers.

CNN reached out to the treasurers from both PACs for comment but did not receive an immediate response.

凯莉·格林利/路透社

据 CNN 审查的录音显示，支持肯·帕克斯顿的超级政治行动委员会“孤星自由”（Lone Star Liberty）从埃隆·马斯克那里获得了 1000 万美元的捐款，该委员会正在德克萨斯州利用人工智能驱动的自动语音电话活动攻击民主党参议员候选人詹姆斯·塔拉里科，指控他从“数据中心特殊利益集团”那里收受资金。

这些此前未被报道过的电话中，一位名叫“詹妮弗”的AI生成的语音来电者自称是“孤星自由党”（Lone Star Liberty）的虚拟工作人员。与传统的预录自动语音电话不同，该系统旨在与选民实时互动，询问他们支持哪些候选人，并对他们的回答作出回应。

“你好，请问是胡里奥吗？你好。我叫詹妮弗，是孤星自由组织的虚拟工作人员，我打电话来支持共和党人肯·帕克斯顿，”CNN 审阅的一段通话录音中，人工智能生成的声音这样说道。

“你好，请问是胡里奥吗？你好。我叫詹妮弗，是一名远程工作人员。”

支持肯·帕克斯顿竞选美国参议院席位的超级政治行动委员会“孤星自由”正在德克萨斯州利用人工智能驱动的自动语音电话活动攻击民主党参议院候选人詹姆斯·塔拉里科。

AI语音播报了共和党人对塔拉里科的攻击性言论——其中一次通话中声称他“从数据中心特殊利益集团那里收受了1100万美元”。

AI 机器人呼叫者在数据中心攻击 James Talarico

德克萨斯州超级政治行动委员会正在使用人工智能工具攻击塔拉里科，因为他接受了来自数据中心的资金，而这些数据中心正是实现此类人工智能决策的前提。

该政治行动委员会发言人表示，这一数字是基于《每日来电》的一篇报道，该报道详细披露了与数据中心行业有关联的个人向支持塔拉里科的政治团体捐款的情况。这些捐款人包括科技亿万富翁里德·霍夫曼，他不顾公众的广泛反对，一直大力支持数据中心扩张，并在本届选举周期中向一个支持塔拉里科的超级政治行动委员会捐赠了1000万美元。

这场自动拨号电话攻击活动呈现出一个惊人的讽刺意味：马斯克，这位世界首富和顶级人工智能公司首席执行官，正花费数亿美元的个人财富支持共和党的政治竞选和超级政治行动委员会（Super PAC）。而在这个案例中，德克萨斯州的超级政治行动委员会正利用人工智能工具，攻击那些为这类电话提供技术支持的数据中心，试图从中牟利。

人工智能自动语音电话并非政治竞选活动所独有——保险和金融等其他行业也在采用它们——但这些电话也让我们得以初步了解竞选活动如何利用对话式人工智能电话进行大规模的政治游说——这是一种新兴的政治宣传形式，公众难以追踪，部分原因是它缺乏集中式数据库。

CNN的KFile栏目通过一款名为Nomorobo的应用程序识别出了这些骚扰电话。Nomorobo是一款可以拦截骚扰电话和短信的应用程序。该栏目获得了通话录音。据Nomorobo称，这些骚扰电话最早于10月23日被检测到，来自数十个不同的电话号码，但只拨打了相对较少的固定电话。

数据中心已成为2026年大选周期中最具争议的政治目标之一，其能源消耗、成本和快速扩张引发了跨党派的反对。CNN最近的民调显示，超过80%的美国人对人工智能的快速发展表示担忧而非兴奋。

Brandon Bell/Getty Images

尽管选民表达了疑虑，但政治竞选活动仍在接受人工智能生成的政治电视和数字广告，以及人工智能驱动的对话式自动语音电话这一新兴领域。

Lone Star Liberty 的发言人 Gregg Keller 在接受 CNN 采访时为该组织的竞选策略辩护，称其“正在利用一切可用手段”向德克萨斯州选民宣传他们对 Talarico 的批评。

塔拉里科竞选团队的一位发言人在一份声明中告诉 CNN：“肯·帕克斯顿找不到人替他打电话，所以埃隆·马斯克正在启动他的数据中心，打造唯一愿意为他们的竞选散布谎言的志愿者：人工智能机器人。”

斯坦福大学法学院教授、斯坦福大学法学院人工智能计划联合主任内特·珀西利在给 CNN 的一封电子邮件中表示，他对人工智能自动拨号电话会影响政治竞选持怀疑态度，因为人们很少接听电话，如果听到“虚拟工作人员”的声音，他们很可能会挂断电话。

“大多数人都觉得自动语音电话很烦人——无论是传统的录音电话还是现在的AI自动语音电话，”珀西利说。“我们应该区分信息传播中虚假或煽动性言论与传播技术本身。我承认AI是这场运动中最引人注目的新事物，但操纵手段和政治一样古老。”

他补充说，人工智能驱动的自动拨号技术的核心问题是“它是否更便宜、更有效，以及它是否能够大规模地做到这一点”。

虽然目前这类人工智能驱动的自动拨号电话通常仅限于人们的固定电话，但根据联邦通信委员会对目前正在审议的一项请愿书的裁决，它们的使用可能会变得更加普遍。

目前，联邦法律禁止大多数自动拨号电话拨打至无线电话，除非事先获得对方同意。联邦通信委员会（FCC）正在考虑是否放宽对政治自动拨号电话的规定——保守派团体“增长俱乐部”（Club for Growth）寻求此项改革，此举可能允许使用人工或预录语音进行大规模的未经请求的电话拨打，不仅能拨打固定电话，还能在未经事先同意的情况下拨打移动电话。

民主党高层和消费者权益组织反对这项请愿。联邦通信委员会（FCC）可能在10月底前做出裁决。两年前，FCC曾宣布，利用人工智能克隆他人声音的自动语音电话违反现行联邦法律，属于非法行为。此前，在2024年美国总统初选中，曾有人拨打电话冒充乔·拜登总统，敦促新罕布什尔州选民不要投票。

CNN审阅的录音既展现了这项技术的潜力，也暴露了其局限性。人工智能来电者能够对接听者做出反应，但当受到干扰时也会出现故障——停止、重新开始，甚至重复提问或竞选宣传要点。为了保护接听者的隐私，Nomorobo已从CNN审阅的录音中移除了接听者的声音。

如果被问及支持哪位候选人，这位虚拟工作人员会回答：“我没有个人意见，因为我是人工智能虚拟工作人员。我可以解释竞选团队的立场，但最终决定权在你。”

“我没有个人观点，因为我是一个人工智能虚拟员工。”

这款人工智能自动拨号器可以与来电者互动并回答问题，但当被问及对选举的看法时，它会拒绝回答。

CNN 对包含“虚拟工作人员”一词的自动语音电话进行审查后发现，至少还有两个其他政治行动委员会在本轮选举周期中使用了相同的“Jennifer”人工智能语音。

今年早些时候成立的“密苏里承诺”政治行动委员会（Missouri Promise PAC）由“黑钱”（即来自不公开捐赠者的团体的资金）资助，该委员会在今年夏天利用人工智能自动拨号系统询问选民对第五修正案的看法。该修正案将授权立法者逐步取消密苏里州的州所得税，并通过提高销售税来弥补税收缺口。最终，选民在8月份否决了这项提案。

今年 10 月，同一个 AI 自动语音电话发起了另一场活动，与“诚信选民”组织合作。该组织是密苏里州的一个政治行动委员会，于 8 月成立，旨在支持多位共和党候选人竞选该州的席位。

但这两个政治行动委员会的AI自动语音电话录音都没有提到数据中心。

CNN已联系这两个政治行动委员会的财务主管征求意见，但尚未收到回复。