Lee showed how he uses AI. Trump isn't telling

President Donald Trump acknowledged using artificial intelligence aboard Air Force One last month but declined to explain how, while President Lee Jae Myung publicly shared a Google Gemini response in a political dispute. Lee’s disclosure drew criticism from opposition politicians, and his office said he used Gemini to verify legal facts rather than make government decisions. The contrast comes as American politicians’ chatbot use raises questions about accuracy, transparency and accountability.

Trump said, “I use AI. Yes, I do,” and, when asked how, replied, “Well, I don’t want to tell you that.”

Sen. Tommy Tuberville told The Washington Post in August that he consults ChatGPT during long drives to learn about issues such as Medicare and Medicaid.

Rep. Anna Paulina Luna defended her staff’s use of AI after a summary accompanying a defense bill amendment inadvertently included text generated by Anthropic’s Claude.

Carnegie Endowment for International Peace researchers Danaé Metaxa and Alex Engler wrote that large language models’ influence on political discourse is a critical new frontier and noted that chatbots can give different responses to identical questions.

Rep. Na Kyung-won said Gemini warned that justifying investigative agency chiefs’ Cabinet attendance through presidential supervisory authority could threaten their political neutrality, and accused Lee of framing his question to get a favorable response.

Published Oct 24, 2026 2:24 am KST

U.S. President Donald Trump, right, meets with South Korean President Lee Jae Myung at the Oval Office, at the White House, in Washington, D.C., Aug. 25, 2025. Reuters-Yonhap

SEATTLE — President Lee Jae Myung has come under fire for publicly sharing how he used Google's Gemini to support his position in a political dispute. Across the Pacific, U.S. President Donald Trump has taken a different approach: He acknowledges using artificial intelligence (AI) but won't say exactly how.

Trump is hardly shy about AI-generated content. His social media accounts have featured a string of fabricated images and videos, including portrayals of himself as a pope and a king, as well as AI-generated content mocking political opponents.

But when reporters asked Trump about his personal use of AI aboard Air Force One last month, he offered little insight.

"I use AI. Yes, I do," Trump said.

Pressed on how he uses the technology, Trump replied, "Well, I don't want to tell you that."

The exchange illustrates a contrast with Lee, who publicly shared a Gemini-generated response to support his position on whether the heads of investigative agencies should attend Cabinet meetings.

Lee's disclosure sparked criticism from opposition politicians, who questioned his reliance on a chatbot to address legal and constitutional matters. The presidential office maintained that Lee had used Gemini to verify legal facts, not to make government decisions.

While Trump's public embrace of AI-generated political imagery is well documented, his remarks offered no indication of whether he personally consults chatbots on policy or legal questions.

Other American politicians, however, have been more forthcoming about their use of the technology.

Sen. Tommy Tuberville, a Republican from Alabama, told The Washington Post in August that he consults ChatGPT during long drives to learn about issues such as Medicare and Medicaid.

Rep. Anna Paulina Luna, a Florida Republican, defended her staff's use of AI after a summary accompanying a defense bill amendment inadvertently included text generated by Anthropic's Claude.

The growing use of AI in American politics has raised questions about accuracy, transparency and accountability, particularly when elected officials rely on chatbot-generated information in their work.

In an August analysis published by the Carnegie Endowment for International Peace, researchers Danaé Metaxa and Alex Engler highlighted the growing influence of AI chatbots on political discourse and the challenges of evaluating their responses.

"We see LLMs’ impact on political discourse as a critical new frontier of modern politics," they wrote, referring to large language models, the technology behind AI chatbots.

The researchers also noted that AI chatbots can produce different responses even when given identical questions.

The dispute surrounding Lee's use of Gemini also illustrated how different questions about the same political issue can yield conflicting AI-generated responses.

Rep. Na Kyung-won of the main opposition People Power Party said she had also consulted Gemini about the issue and received a response warning that justifying the attendance of investigative agency chiefs through presidential supervisory authority could threaten their political neutrality.

Na accused Lee of framing his question to obtain a favorable response.

上个月，美国总统唐纳德·特朗普承认在空军一号上使用过人工智能，但拒绝解释具体操作方式。与此同时，韩国总统李在明则公开分享了谷歌“双子座”人工智能在一次政治争论中的回应。李在明的公开行为招致反对派政客的批评，其办公室则表示，他使用“双子座”是为了核实法律事实，而非用于政府决策。这一对比引发了人们对美国政客使用聊天机器人的准确性、透明度和问责制的质疑。

特朗普说：“我使用人工智能。是的，我的确使用。”当被问及如何使用时，他回答说：“嗯，我不想告诉你。”

参议员汤米·图伯维尔 (Tommy Tuberville) 在 8 月份告诉《华盛顿邮报》，他在长途驾驶期间会咨询 ChatGPT，以了解医疗保险和医疗补助等问题。

众议员安娜·保利娜·卢娜为她的工作人员使用人工智能辩护，此前一份国防法案修正案的摘要无意中包含了由 Anthropic 公司的 Claude 生成的文本。

卡内基国际和平基金会的研究人员达娜·梅塔克萨和亚历克斯·恩格勒写道，大型语言模型对政治话语的影响是一个重要的新领域，并指出聊天机器人可以对相同的问题给出不同的回答。

罗京元议员表示，Gemini 警告称，通过总统监督权来证明调查机构负责人出席内阁的合理性可能会威胁到他们的政治中立性，并指责李在提问时故意设置问题以获得有利的答复。

发布于2026年10月24日凌晨2:24（韩国标准时间）

2025年8月25日，美国总统唐纳德·特朗普（右）在华盛顿白宫椭圆形办公室会见韩国总统李在明。（路透社-韩联社）

西雅图——韩国总统李在明因公开分享其如何利用谷歌的Gemini项目来支持其在政治争端中的立场而受到抨击。而在太平洋彼岸，美国总统唐纳德·特朗普则采取了不同的做法：他承认使用了人工智能（AI），但拒绝透露具体细节。

特朗普对人工智能生成的内容毫不避讳。他的社交媒体账号上充斥着大量伪造的图片和视频，包括将自己描绘成教皇和国王，以及利用人工智能生成内容嘲讽政治对手。

但当记者问及特朗普上个月在空军一号上亲自使用人工智能的情况时，他却没有提供太多信息。

“我使用人工智能。是的，我的确使用，”特朗普说。

当被追问如何使用这项技术时，特朗普回答说：“嗯，我不想告诉你。”

这段交流与李显龙形成了鲜明对比，他曾公开分享了Gemini生成的回复，以支持他关于调查机构负责人是否应该参加内阁会议的立场。

李显龙的披露引发了反对派政客的批评，他们质疑他依赖聊天机器人处理法律和宪法事务的做法。总统府则坚称，李显龙使用Gemini是为了核实法律事实，而不是为了做出政府决策。

虽然特朗普公开接受人工智能生成的政治图像已是众所周知的事实，但他的言论并未表明他是否会就政策或法律问题亲自咨询聊天机器人。

然而，其他一些美国政治家则更愿意公开谈论他们使用这项技术的情况。

阿拉巴马州共和党参议员汤米·图伯维尔 (Tommy Tuberville) 在 8 月份告诉《华盛顿邮报》，他在长途驾驶期间会咨询 ChatGPT，以了解医疗保险和医疗补助等问题。

佛罗里达州共和党众议员安娜·保利娜·卢娜为她的工作人员使用人工智能辩护，此前一份国防法案修正案的摘要无意中包含了由 Anthropic 公司的 Claude 生成的文本。

人工智能在美国政治领域的日益普及引发了人们对准确性、透明度和问责制的质疑，尤其是在民选官员依赖聊天机器人生成的信息开展工作时。

卡内基国际和平基金会在 8 月份发表的一份分析报告中，研究人员 Danaé Metaxa 和 Alex Engler 重点介绍了人工智能聊天机器人对政治话语日益增长的影响，以及评估其回应所面临的挑战。

他们写道：“我们认为 LLM 对政治话语的影响是现代政治的一个关键新前沿”，这里指的是大型语言模型，即人工智能聊天机器人背后的技术。

研究人员还指出，即使面对相同的问题，人工智能聊天机器人也可能产生不同的回答。

围绕李使用 Gemini 的争议也表明，针对同一政治问题的不同问题可能会产生相互矛盾的 AI 生成答案。

主要反对党国民力量党议员罗京媛表示，她也曾就此问题咨询过双子座，得到的回复警告称，通过总统监督权来证明调查机构负责人出席的合理性可能会威胁到他们的政治中立性。

Na指责Lee故意设置问题以求获得有利的回答。