New Zhiyuan reports
This is absolutely terrifying.
Just now, Trump revealed—
that he had once threatened to arrest the CEO of an AI giant because he had used AI to do some bad things and might have done so maliciously.
As for whom he was referring to, the answer is practically obvious.
At the beginning of the year, Trump threatened that he would face “criminal consequences.”
In September, he publicly named him, saying, “I’ve stopped him from doing bad things.”
In addition, Trump also once admitted privately, “A week earlier, I still considered him a national security threat.”
Trump’s attitude toward the company has been like a roller coaster.
What exactly happened?
The February trigger: a conflict set off by a “refusal”
In February 2026, Anthropic’s Claude had already been used for classified operations.
The powers that be wanted to further relax the restrictions—allowing AI to be used for lethal weapons or mass surveillance.
But Amodei refused.
He insisted on the company’s “red lines”: it must never be used for lethal weapons without human oversight, nor for mass surveillance.
This time, Trump was absolutely furious.
Around February 27, Trump posted directly on Truth Social, taking aim at Anthropic.
I have directed every federal agency of the US government to immediately stop using all Anthropic technology!
We don’t need it, we don’t want it, and we will never do business with them again!
He gave them a six-month transition period and issued a stern warning:
Anthropic had better wise up and cooperate during the transition period. Otherwise, I will use the full power of the presidency to make them comply, and major civil and criminal consequences will follow.
Good lord, major criminal consequences. That sounds terrifying.
The Pentagon promptly listed Anthropic as a “supply chain risk.”
Previously, this label was reserved for competitors.
Anthropic’s name on the list truly is a first in history.
Later, Amodei quipped in an internal memo—
“The government doesn’t like us because we didn’t donate money to Trump or lavish him with ‘dictatorial praise’ like other companies did.”
That left problems in its wake.
Later, Anthropic explicitly warned investors in its IPO filing that the company could face targeted retaliation from the government because of its AI safety stance.
Why does Trump regard AI safety as a “scam”?
Why would Trump strike so hard against a company that advocates “AI safety”?
In his latest interview on 60 Minutes, Trump gave the answer himself.
He showed them absolutely no mercy, mercilessly mocking Silicon Valley’s elite.
During the interview, the host asked: “Will AI pose a threat to humanity’s future?”
Trump bluntly said that, in the world today, only two sides are engaged in the real SI race.
“I brought all those AI moguls to the White House,” Trump said derisively.
The host was thoroughly confused: “These people who work on AI are so bizarre, so strange. They even stand outside their own companies and say, ‘Hey, we created this incredible company, but it has a 10% chance of killing all of humanity!’”
“It’s like standing in front of a new hotel and shouting to everyone, ‘This is my new hotel, but it might collapse!’”
In Trump’s eyes, the so-called “AI destroys humanity” is a “scam” obstructing his way forward.
With his next sentence, he instantly punctured the illusion of safety.
“You ask me why I’m not worried about doomsday panic? Because other people will have this technology! If we put in the so-called guardrails, we’ll be at an absolute disadvantage. Other people won’t have those guardrails… If AI is dangerous, then it’ll be winner-take-all!”
Amodei: From “Perfect Little Angel” to “Sinister Plot”
As early as September this year, the AI industry was swept up in a “slowdown wave.”
Amodei published a lengthy article openly calling on major laboratories to slow the development of the most advanced AI so that safety measures could keep up.
Even more strikingly, OpenAI’s Altman and Musk both supported the call.
Trump blew up again
On September 14, he unleashed a barrage of posts.
“The Trump administration has already stopped those AI ‘people’ from doing bad things, or possibly ‘things’ that might be bad, such as Dario (Anthropic!), who now pretends to be a ‘perfect little angel’—and we will continue to do so!”
“We already have enormous criminal and regulatory powers over these companies!”
“A sinister conspiracy is underway against AI and data centers, and the only ones happy about it are the competitors. Whoever wins in AI wins! Conspiracy theorists, traitors, turncoats, leakers, beware!”
In Trump’s view, AI is the key to the United States winning the future, and any slowdown would be bad for America.
In his eyes, the real “guardrail” is “a strong and intelligent (high-IQ!) American president.”
Interestingly, the two did not end in a rupture.
At the end of September, Trump even had a private dinner with Amodei.
Afterward, Trump told the media that he liked Amodei and his wife, found them “very smart,” and thought they were not quite what he had expected.
The two later also participated together in the White House-organized signing of the “Superintelligence Safety Agreement.”
Their relationship seemed to have returned to a delicate balance.
However, when the most powerful person in the world can publicly imply that “keep playing with fire and you’ll go to prison,” AI is no longer a technological issue, but a question of power.
It’s all about power.
References:
https://www.reddit.com/r/singularity/comments/1xrhqeh/trump_claims_that_he_threatened_to_arrest_a_major/https://www.youtube.com/watch?v=JklRXlv69t4
Editor: Aeneas David
Instantly follow ASI
⭐Like, share, and recommend—all with one click!⭐
Light up the star and never miss New Zhiyuan’s lightning-fast updates!
