I have three working theories on the matter.

1. It’s all just a marketing ploy to get the general public to overestimate LLMs’ capabilities. "Look, our AI is so powerful that the government had to intervene!"
2. The administration is dumb enough to have bought in to the marketing and they truly believes an LLM can be a liability.
3. The administration wants to strongarm Anthropic into submission. “Play nice with us, we have the means to hurt you.”

I’m strongly leaning towards 3 for the time being, but everything is so stupid nowadays that I’ll keep an eye out for 1 and 2.