Anthropic chief executive Dario Amodei attended a White House dinner with President Donald Trump on Sunday after critics circulated a memo attacking his political record and questioning his trust in the president.
The memo reportedly accused Mr Amodei of having “a long record of attacking Trump” and “deep Democratic ties”. It also claimed that “he doesn’t trust President Trump”.
The episode highlights the political pressure facing technology executives seeking access to the White House, where agreeing to meet the president does not necessarily shield them from criticism by his allies.
White House seeks to bring AI chiefs together
It was the second recent memo from Republicans targeting leaders in the artificial intelligence industry. An earlier document, titled “Effective Altruism Opposition Research File”, attacked people in the sector who have raised concerns about AI safety.
That document characterised such figures as effective altruists, left-wing, vegan and polyamorous, and made a claim about their attitudes towards sex with animals.
The attacks come as Anthropic and other developers of advanced AI models have disclosed incidents involving systems and agents escaping sandboxes designed to isolate them from the internet. The companies have also reported models hacking multiple businesses and government websites.
Anthropic, along with other leading AI firms, has suggested establishing an international mechanism to review models for flaws before they are released.
However, the Trump administration does not want to slow AI development, viewing it as a race against China that the United States must win. Anthropic and OpenAI are also heading towards potential initial public offerings valued at more than $1 trillion, making their leaders difficult for the White House to overlook.
“The WH is trying to get everyone seated at the table,” a source said.
