Artificial intelligence is increasingly being used for complex technical tasks, but a new threat intelligence report has raised serious concerns over its potential misuse in weapons development, cyberattacks, mass surveillance and online fraud.
The report details suspected operations linked to China, Russia, Iran, Yemen, Mali and Europe involving the Claude AI model. The cases have been presented as findings and allegations from the investigation and have not been independently verified in every instance.
Was AI Used for Weapons?
According to the report, a China-linked operator used Claude to develop software related to electronic warfare and suppression of air-defence systems.
The project allegedly included prioritising potential targets and modelling radar jamming. A simulation included 12 targets in Taiwan, among them early-warning radars, Patriot and Tien Kung missile systems, air bases and a command bunker.
Accounts associated with the activity were subsequently banned.
Was Claude Used for Missile Software?
A suspected Yemen-linked group allegedly used Claude while working on software for a guided rocket, a proposed ballistic missile with a range exceeding 2,000 km and a missile variant incorporating a hypersonic glide vehicle.
The reported work included coding, simulations and troubleshooting.
However, the report found no evidence that these efforts resulted in a successfully operational weapon.
Russia-linked operators were also accused of using AI to develop software for autonomous swarms of first-person-view attack drones, including target selection, terminal guidance and coordination between multiple aircraft.
What Other Military Uses Were Found?
In another China-linked case involving the Chinese navy, Claude was allegedly used to develop technical specifications and fire-control software for an anti-torpedo system.
The AI reportedly helped prepare a technical proposal exceeding 200 pages and compare the proposed system with US Navy technology.
A separate China-based defence intelligence operator reportedly used AI to research high-power microwave weapons, identify components and suppliers and trace relevant supply chains.
How Was AI Used in Cyberattacks?
Chinese-speaking operators potentially based in Changsha allegedly targeted around 50 organisations, including foreign government networks.
According to the report, AI-driven workflows were used to identify previously unknown software vulnerabilities, develop methods to exploit them and conduct parts of cyber intrusions with limited human supervision.
A Russia-linked group was separately accused of targeting Ukrainian government, military and diplomatic sectors through activities including phishing, hotel Wi-Fi intrusion and WhatsApp account takeovers.
AI was allegedly used at several stages of these operations.
Was AI Used for Military Intelligence?
An Iran-linked operator allegedly used Claude to gather information that could support potential targeting of US naval forces.
The reported activity involved analysing publicly available photographs, ship and aircraft identification information and commercial satellite imagery.
The operator also reportedly researched vulnerabilities affecting communications equipment and industrial control systems.
Proposal for Conducting Cyber Crisis Drill, Tabletop Exercise (TTEx) & CCMP Readiness Exercise
How Was AI Used for Surveillance?
The report describes several alleged surveillance operations.
In one China-linked case, Claude was allegedly used to identify and approach members of the Uyghur community in Syria, including people considered financially or personally vulnerable.
Other alleged operations involved gathering information on Catholic religious leaders, Taiwanese Christian leaders, Tibetan Buddhists, dissidents and activists and preparing surveillance reports.
In Mali, Claude was reportedly used as an engineering tool in the development of a domestic surveillance system capable of monitoring data associated with around 25 million SIM cards across the country’s three mobile networks.
Were Political Groups Targeted?
A French-speaking operator reportedly targeted 42 European political parties, media organisations, think tanks and software service providers used by them.
The campaign allegedly involved stealing data and login credentials and developing a system for collecting and organising personal information about individuals.
The report presents this as another example of AI being incorporated into operations involving political and personal information.
How Were Dating Apps Used?
One of the largest alleged fraud operations described in the report involved more than 20 dating apps developed by a China-linked operation.
The apps reportedly contained more than 4,700 AI-generated profiles that interacted with at least 25,000 users.
According to the report, automated AI personas were combined with paid human workers to make the profiles appear genuine and maintain prolonged conversations with users.
What About Biological Research?
The report also describes five cases involving attempts to use AI in ways that could support biological weapons development.
These reportedly included attempts to obtain assistance for research involving modification of the chikungunya virus and efforts to bypass service restrictions while researching highly pathogenic avian influenza.
Other cases involved research related to orthopoxviruses and novel venoms and toxins.
Did Claude’s Safeguards Stop Them?
The AI company said its safeguards blocked many of the reported attempts and that action was taken against accounts associated with suspicious activities.
However, some operations allegedly managed to bypass safeguards.
China has said it was not aware of the allegations and maintained that artificial intelligence should be developed for beneficial purposes.
The420 AI Watch: When AI Becomes a Weapon
The cases highlight a growing challenge for AI companies and governments. Powerful AI systems can make sophisticated technical work easier and faster, but the same capabilities can potentially be exploited for cybercrime, surveillance, fraud and other harmful activities. Stronger safeguards, faster detection of misuse and continued monitoring will become increasingly important as AI systems grow more capable.
About the author — Ayesha Aayat writes on cybercrime, digital safety, and emerging online threats. Her work focuses on public awareness, legal clarity, and technology-driven risks.
Follow for daily updates on cybercrime, corporate fraud, DFIR, hacking, investigations, and digital forensics
