Anthropic, the developer behind the Claude AI chatbot, has come under fresh scrutiny following a report detailing how its security officials engage in monitoring and surveillance, including reporting AI activists to police over potential future crimes. The claims, based on job postings and interviews with senior security personnel, describe a company tracking individuals who oppose the rapid development of artificial intelligence near its physical premises.
The report describes how Anthropic monitors activists positioned close to what it terms “physical Anthropic assets”. In an earlier statement, the company acknowledged making routine reports to police departments regarding threats against employees, and said it monitored behaviour in an effort to identify escalation patterns at an early stage.
Anthropic has disputed the framing of the coverage. In an emailed response, the company said the headline and characterisation were inaccurate, stating: “Anthropic is not — and has never said it is — ‘building a predictive surveillance system to monitor activists.’ This is a physical-security function focused on keeping our employees safe, the same as any large company with a global workforce.”
A Job Posting at the Centre of the Debate
Much of the attention centres on a recent Anthropic vacancy for an “enterprise intelligence specialist”. The role is described as identifying, tracking and investigating “global threats including geopolitical instability, terrorism, crime, activism, nation-state targeting of the AI sector, and emerging security trends”.
Anthropic did not dispute the wording of the posting but said the inclusion of “activism” was simply “one of many threat categories a security team monitors for employee-safety purposes”.
The controversy arrives amid an ongoing standoff between Anthropic and the Pentagon, in which the company has clashed with the military over the deployment of mass-surveillance tools and fully autonomous weapons. Anthropic lost several government contracts during the dispute. More recently, after the US Secretary of Commerce, Howard Lutnick, said the company was “back on the right side”, a Pentagon official rejected the claim and described Anthropic as a “supply chain risk”.
Predictive Policing and the Rise of “Precrime”
The developments coincide with growing public backlash over artificial intelligence, particularly concerning data centres and automated licence plate readers. Alongside this, the practice of predictive policing continues to expand. Often referred to as “precrime” in science fiction, this form of crime forecasting has become increasingly integrated into police work, with algorithmic software and AI tools analysing historical crime data, arrest records and individual “risk scores” to pre-emptively flag criminal activity.
Anthropic has access to substantial volumes of data through its Claude AI chatbot and other intelligence-gathering tools. According to reporting, the company has referred individuals to police while declining to share evidence of threats or wrongdoing. These accounts allege that Anthropic could be bypassing standard legal protocols, both by reporting alleged “precrimes” without due process and by withholding evidence from law enforcement.
Anthropic maintains that the function relates solely to physical security and the safety of its workforce.
