Trump’s administration has sacked Collin Burns from his job as head of the Centre for AI Standards and Innovation (CAISI) just four days after he was appointed. The expert, who used to work at OpenAI and Anthropic, got caught up in a political spat: the White House isn’t happy about his links to Anthropic, which the Trump administration has officially labelled as a threat to national security. Even though Burns turned down millions in stock options to focus on public service, he was replaced by Chris Fall, a scientist who specialises in systemic problems. This whole thing shows just how much the MAGA administration and the people making super-powerful models like Mythos don’t get on.
The Donald Trump administration has sacked Collin Burns, an artificial intelligence specialist who was just appointed as head of the Centre for AI Standards and Innovation (CAISI) by the US Department of Commerce. The Washington Post is reporting this, and they’ve got their sources.
Burns, who’d previously worked at Anthropic and OpenAI, started at the centre on Monday 20 April, but by Thursday 23 April the White House had insisted he resign. The reason was concerns over his links to Anthropic, a company that’s been in a bit of a row with the Trump administration recently.
I’ve heard from a couple of people who’ve spoken to the Washington Post that some of the higher-ups in the White House weren’t told about Burns’ appointment beforehand. They were worried about the idea of someone from a big private AI lab having a role that involves working closely with the industry.
CAISI, which was set up during Joe Biden’s time as vice president and later renamed under Trump, is the government’s main way of working with the AI industry. The centre looks at the national security risks of the latest models, especially the Anthropic Mythos, which the company says could be dangerous when it comes to cyber hacking. With these models growing in power, the administration is trying to keep control while also staying connected to the people who matter most in the market.
The right-wing Daily Signal was the first to report on Burns’ appointment, and the reaction from the expert community was pretty positive. Zvi Movshovitz, the well-known AI blogger, has said in public that he thinks the decision to appoint him was an “excellent choice”. But then, a few hours later, the publication updated its article, saying that the department had “decided to move in a different direction” and picked a new candidate.
Chris Fall is the new head of the centre. He’s a scientist with tons of experience in the government and at universities. The Department of Commerce said that Fall would provide “the scientific leadership necessary for America to lead the world in evaluating cutting-edge AI models and developing technical standards that protect national and economic security”. The White House passed all questions about the situation to the Department of Commerce and didn’t say anything about why they’d changed candidates.
What happened in Burns’s past makes the scandal even more dramatic. He was studying for a PhD in computer science at the University of California, but he left the programme in 2023 to join OpenAI, and then he moved to Anthropic at the end of 2024. When he was younger, he set a world record for solving a Rubik’s Cube in 5.25 seconds.
Dean Ball, who used to be a Trump administration adviser on AI, said on social media that Burns had given up valuable shares in Anthropic and moved across the country to work for the government, only to get “a slap in the face from his own government” in return. He said that Burns’ resignation was linked to his connection with Anthropic, calling it a “silly but predictable own goal” by the administration.
But there are also signs that relations between the White House and the company are improving. The Mythos model has made the government sit up and take cybersecurity risks associated with new AI systems seriously, and recently the White House held talks with Anthropic CEO Dario Amodei, where they discussed cooperation. But the parties are still fighting it out in court over the Trump administration’s decision to label the company as a threat to national security.