Trump keeps a close eye on AI: One week ago Anthropic was considered a "national security threat", and now does not rule out intervention when necessary.

President Trump's attitude towards AI star start-up Anthropic seems to have changed, but this does not mean that the White House has relaxed its vigilance on cutting-edge artificial intelligence. According to a report on Friday, the 19th, President Trump stated in an interview with American media that he currently does not consider Anthropic and its CEO Dario Amodei to pose a national security threat, but admitted that he may have thought so "a week ago". Trump mentioned that he met with Amodei during the G7 summit in France and acknowledged Amodei's prompt response to government concerns. However, when asked if he would shut down Anthropic, Trump first said he would not shut it down, but later said, "I would do it, but I'm not sure if I have to." Furthermore, in the interview, Trump did not rule out the possibility of using the powers granted by the Defense Production Act (DPA) to force Anthropic to cooperate with the government. This statement indicates that although the White House has temporarily eased its stance, it has not ruled out the possibility of directly intervening in Anthropic and the entire AI industry in the future. Analysts believe that the rapid change in attitude towards Anthropic within a week indicates that the Trump administration is re-evaluating the AI industry from a national security perspective and trying to find a balance between maintaining America's leading edge and mitigating the risk of technology proliferation. From "National Security Threat" to "Very Responsible" at Anthropic When asked if he thought Anthropic and Amodei posed a threat to U.S. security, Trump replied, "Not now. But maybe a week ago." Trump revealed that he met with Amodei and other tech industry executives during the G7 summit and believed that Anthropic quickly responded to U.S. government concerns. "He quickly responded to us because it involves tremendous responsibility," Trump said. However, Trump did not completely dispel doubts. He stated that he would not shut down Anthropic, saying, "I would do it, but I'm not sure if I have to. I think so far, they have been very responsible." This means that while the White House has not taken further actions for now, it still retains the option to intervene when necessary. U.S. Government Imposes the Strongest Regulation on AI Companies to Date Just a few days ago, the U.S. Department of Commerce required Anthropic to obtain U.S. government approval before allowing foreign individuals, companies, or even foreign governments to use its advanced Fable 5 and Mythos 5 models. This move is seen as the most significant intervention by the U.S. government on the operations of an artificial intelligence company to date. Prior reports suggested that the Trump administration was concerned that advanced AI models could become strategic resources similar to nuclear technology and advanced chips, necessitating restrictions on their overseas dissemination. U.S. Secretary of Commerce Lutnick previously wrote to Anthropic CEO Dario Amodei, requesting the company to strengthen control over the use of the models. Subsequently, senior executives and technical leaders of Anthropic quickly engaged in discussions with Trump administration officials to alleviate concerns about the model's capabilities and international proliferation risks. In the interview released on Friday, Trump said, "We didn't like some of the things they did before, but they responded very quickly." He specifically praised Amodei as "very smart" and stated that Anthropic has been performing well so far. AI Competition a Priority, But Government Retains the "Power to Act" Trump emphasized that the U.S. is leading the AI competition against other countries and therefore will not easily restrict the development of domestic companies in the U.S. However, this does not mean that the government will abandon regulation. Trump believes that the AI field involves "tremendous responsibility," and the government must ensure that the technology remains within controllable boundaries. When discussing the Defense Production Act (DPA), Trump said, "I have a lot of means I can use," but "I'm not sure if I have to use it." Regarding concerns raised by Amazon, Trump said, "Amazon, as both a competitor and a part-owner, reported Anthropic. They are dissatisfied with Anthropic's practices and have deep concerns. I think so far, all parties have been very responsible in handling the situation." Analysts believe that the change in the Trump administration's attitude towards Anthropic does not mean relaxation of regulation but sends a clearer signal: the U.S. will support the development of companies as long as they are willing to comply with government requirements. However, if the White House deems that cutting-edge models may threaten national security, the government will not hesitate to wield its administrative power for intervention. Critical Stage for Anthropic IPO Provides Breathing Room For Anthropic, the easing of Trump's attitude comes at the right time. Bloomberg reported that Anthropic submitted IPO application documents weeks ago, with the company's latest valuation exceeding $90 billion. Previously, the restrictions imposed by the U.S. Department of Commerce had raised market concerns, suggesting that if the White House continued to view Anthropic as a national security risk, the company's commercial prospects and listing plans could be impacted. Now, with Trump publicly stating that he does not believe Anthropic poses a threat, it has temporarily eased external concerns. However, his statement of being prepared to take action when necessary also means that the regulatory sword hanging over the AI industry has not disappeared. As the U.S. government increasingly views the most advanced AI models as national strategic assets, future AI giants such as OpenAI, Anthropic, Google, etc., may face a regulatory framework similar to that of the semiconductor industry. For the Trump administration, ensuring that the U.S. wins the AI race is important, but preventing technology proliferation and controlling potential risks are also becoming equally critical strategic objectives. This article is translated from "Wall Street sees news", author: Li Dan, GMTEight, edited by Li Cheng.