AI makes it "easier than ever" for adversaries to target U.S. military, experts say

Anthropic said that a group it did not identify, which it labeled a threat actor, used its Claude AI model to compile open-source information to identify U.S. naval positions and to develop targeting recommendations. The company said it banned the account. Anthropic also said it shared the information with the U.S. government in order to disrupt the threat. military, experts say.'

According to the excerpt, the group used Claude to gather and organise publicly available, or open-source, material, and to produce targeting recommendations based on it. The excerpt states that Anthropic banned the account and that it shared its findings with the U.S. government to disrupt the threat. The excerpt does not name the group, describe its location or affiliation, or specify which part of the U.S. government received the information. It also does not state when the activity occurred, how long it lasted, or how the account was identified.

The excerpt attributes the broader claim — that AI makes it 'easier than ever' for adversaries to target the U.S. military — to experts cited in the article, rather than to Anthropic. Anthropic's account of the episode is the source given for the specific details: the use of Claude, the compilation of open-source information, the identification of naval positions, the targeting recommendations, the account ban, and the notification of the U.S. government. The excerpt does not include comment from the U.S. government about the information it received.