A government-linked consultant used Claude to help build a system capable of intercepting mobile communications across an entire country and creating detailed information on people being watched, Anthropic has said in a new report on the misuse of AI.Anthropic said the case involved a single consultant working for Malian national security authorities. According to the company, the consultant used Claude to engineer a mass-interception platform that could surveil communications across all of the country's mobile operators and generate information on targets.The details were published in Anthropic's September 2026 report titled “Detecting and countering misuse of AI”. The company said it identified and disrupted several surveillance operations between January and July involving state-aligned actors, state-linked contractors and commercial surveillance firms.Claude was used to build the surveillance systemAnthropic said the Mali case was different from situations where an AI tool is simply used to analyse information that has already been collected.According to the company, Claude was used to engineer the software that enabled the intelligence gathering itself. Anthropic described the case as an example of AI being used in place of a larger engineering workforce.The company said a single consultant was able to use Claude to help develop a system designed to intercept communications across Mali's mobile operators.Anthropic said the platform was also capable of generating detailed information on targets. The company did not say that Claude itself independently carried out the interception. Instead, its report said the AI assistant was used to build the underlying software that enabled the surveillance.AI was also used to sort information on peopleAnthropic said the surveillance operations it investigated went beyond building technical tools.In one case, an actor uploaded batches of social media posts and asked Claude to turn them into structured records. According to Anthropic, the records included information such as people's locations, demographic details and political views, along with confidence scores.The company said this allowed large amounts of online information to be processed and organised around individual targets.Anthropic also said an Iranian unit used Claude to analyse hundreds of thousands of social media posts and selected 39 opposition accounts for monitoring.In another case, actors in China used Claude to score social media content and news articles based on political sensitivity and flag possible targets for what they called “control”, according to the company.AI was used in other surveillance operationsAnthropic said it also identified a PRC-aligned operation targeting Uyghurs in Syria.According to the company, the actor involved did not have Arabic language skills but used Claude during a multiday recruitment operation. Anthropic said Claude drafted messages in a regional dialect, translated replies in real time and role-played as an expert to check the quality of the operation.The company said Claude also formatted the results for what Anthropic suspected was a handoff to a case officer.Anthropic described another case in which a Chinese state security bureau used Claude to prepare an internal manual on using AI in surveillance operations.The company said two Iranian units also independently used Claude to solve technical and usability problems in a central surveillance case-management system. Anthropic said the units had no shared code or personnel.Anthropic banned accounts linked to the operationsAnthropic said the operations described in its report violated its Usage Policy, which prohibits non-consensual surveillance and profiling and the use of its services to violate civil liberties and human rights.The company said it banned the accounts associated with the activity and improved its ability to detect the techniques used by the operators.Where the activity involved effects outside Anthropic's platform, the company said it shared identifiers and intelligence with industry partners and authorities where appropriate.Anthropic said that in nearly every case described in the surveillance section, the operators were state-aligned organisations targeting communities that these governments had historically targeted.The groups mentioned by the company included pro-democracy figures in Hong Kong, Tibetan and Falun Gong communities across Asia, as well as Iranian minority communities and opponents of the Iranian government living abroad.The report highlights Anthropic's concern that AI is increasingly being used not only to analyse information but also to build the systems used to collect it, process information about large numbers of people and identify those who may be monitored further.