Will's Blog: Anthropic says it worked with the US National Nuclear Security Administration on a classifier to block concerning chats, like those on building nuclear reactors (Morgan Chalfant/Semafor)

21 August 2025

Anthropic says it worked with the US National Nuclear Security Administration on a classifier to block concerning chats, like those on building nuclear reactors (Morgan Chalfant/Semafor)

Morgan Chalfant / Semafor:
Anthropic says it worked with the US National Nuclear Security Administration on a classifier to block concerning chats, like those on building nuclear reactors — With the government's help, Anthropic built a tool designed to prevent its AI models from being used to make nuclear weapons.

Posted from: this blog via Microsoft Power Automate.