Anthropic's Claude Artificial Intelligence (AI) model has been used in the US military campaign to arrest Venezuelan President Nicolás Maduro.
The information was published by the US newspaper Wall Street Journal, citing many well-informed sources.
The implementation of Claude takes place through a cooperation relationship between Anthropic and data company Palantir Technologies - a unit with software platforms widely used by the US Department of Defense and federal law enforcement agencies.
The Pentagon is promoting leading AI companies - including OpenAI and Anthropic - to bring artificial intelligence tools into secure networks without applying many standard restrictions intended for ordinary users.
Many AI companies are building their own tools for the US military, but most only operate on unsecure networks, often used for military administrative work.
Anthropic is the only company with tools deployed in a secure environment through a third party, but the government still has to comply with the company's usage policies.
Anthropic's usage policy - a company that has just raised 30 billion USD in the latest funding round and is valued at about 380 billion USD - prohibits using Claude to support violence, weapon design or conduct surveillance activities.
The US arrested Venezuelan President Nicolás Maduro in a raid in early January, then took him to New York. Maduro is currently facing drug trafficking charges.
