Anthropic, a prominent artificial intelligence company led by Dario Amodei, has partnered with an external firm, Samdesk, to monitor and identify individuals or groups that may pose a risk to its operations or executives. According to an investigation by The American Prospect, Anthropic regularly shares information about users it deems threatening with law enforcement. This collaboration, managed by Anthropic’s security team—led by Keon Ellison and Zach Melvin—helped the company avoid a planned demonstration during one of its executives’ travels. Ellison mentioned in a podcast that an alert from Samdesk provided early warning, allowing them to divert an executive from the event. Anthropic also has an internal system designed to identify individuals it considers high-risk before involving law enforcement. Public skepticism of AI has been growing in the U.S., and Amodei has publicly expressed concern about the negative perception of AI. The company’s website lists a position for an "Enterprise Intelligence Specialist," offering a salary between $180,000 and $230,000 per year. The role requires identifying threats related to activism, terrorism, and industrial espionage. The Wall Street Journal reported in July that an internal process had flagged several individuals before they were reported to police. A security official from Anthropic said the goal is to gather intelligence that is proactive, predictive, and preventive. In August 2023, a man reportedly told Anthropic’s AI assistant, Claude, that he had just purchased an AR-15 rifle and had Amodei in his sights. Five days later, Anthropic alerted San Francisco police. However, according to a police report reviewed by the San Francisco Standard, an Anthropic employee refused to provide investigators with the full message, citing internal policy. The report did not result in an arrest, and the man later claimed he was joking. Anthropic recently renewed military contracts after a dispute with the Pentagon. A report published by the company acknowledged that AI tools had been misused by groups from countries like Russia, China, Iran, and Yemen for tasks such as missile design and espionage. In Mali, a consultant reportedly used Anthropic’s AI to create a surveillance system tracking 25 million SIM cards without proper legal oversight. Earlier in 2023, Anthropic had refused to allow its AI models to be used for internal surveillance or autonomous weapons, prompting the Pentagon to exclude the company from its contracts. However, a California judge later overturned this decision, stating that a company critical of the government should not be excluded from contracts under the guise of national security. According to The American Prospect, Anthropic is now seeking to re-enter the military sector by hiring for roles related to national security. Journalist Daniel Boguslaw suggests this shift is connected to the political goals of "Americans for Responsible Innovation," a group that wants AI to be classified as critical infrastructure. This classification would require AI companies to report cybersecurity incidents to federal authorities and could expand their access to intelligence from government agencies. Boguslaw warns that this move might lead to a bipartisan effort to label certain data center-related activities as extremism or terrorism, depending on how AI is regulated in the future.