Trump Suddenly Changes His Mind, Long-Time Enemy Becomes Ally Again
US President Donald Trump has suddenly changed his mind again. Previously, Trump had taken a hard stance prohibiting government agencies from using artificial intelligence (AI) tools from Anthropic.
The reason was simple: Anthropic refused to allow its AI tools to be used by the US military for creating autonomous weapons or spying on US citizens. Trump immediately labelled Anthropic as a left-wing ‘woke’ company that endangered the lives of US citizens.
Even the Department of Defense/War (Pentagon) did not hesitate to include Anthropic on the ‘supply chain risk’ blacklist. This incident occurred alongside the joint US and Israel attack on Iran.
According to a WSJ report a short while ago, the Trump administration was actually still using Anthropic’s AI tools in the war against Iran. Those tools were used for combat simulations and enemy detection.
Several months later, Trump has now changed course and reopened discussions with Anthropic. On Tuesday (21/4) local time, Trump stated that Anthropic was “developing well” in the eyes of his administration.
That comment has reopened the path to cancelling the Anthropic boycott at the Pentagon.
Quoted from Reuters on Wednesday (22/4/2026), Anthropic CEO Dario Amodei met with White House officials to improve inter-institutional relations. The White House described the meeting as productive and constructive.
“They [Anthropic] came to the White House a few days ago, and we had a good discussion,” Trump said on CNBC International’s ‘Squawk Box’ programme on Tuesday (21/4).
“I think they’re developing well. They’re very smart and I think they can be useful. I like smart people. I think we’ll get along well with them,” Trump added.
When asked if there were any plans for a new agreement between Anthropic and the Pentagon, Trump said he did not rule out the possibility. “We want the smartest people,” Trump stated.
Anthropic also described their meeting with the White House as going well. The company stated that the discussion focused on collaboration between the two parties on priorities such as cybersecurity, US leadership in the AI race, and ensuring AI safety.
These discussions took place just a few weeks after Anthropic released Mythos, its latest and most advanced AI model, which has banks around the world on edge. The model is said to be able to easily detect hidden vulnerabilities in systems, which could also facilitate exploitation if used by irresponsible parties.
Anthropic assured that Mythos would not yet be widely available. Instead, the company announced Project Glasswing, in which they invited major technology companies, cybersecurity vendors, and US bank JPMorgan Chase, along with dozens of other organisations, to privately evaluate the model and prepare appropriate defences.
Anthropic co-founder Jack Clark said last week that his company was in discussions with the Trump administration regarding the Mythos AI model. However, the details of those discussions were not explained in depth.