Russia Exploits US-Built AI for Drones and Disinformation Campaigns
Kremlin-backed actors have reportedly harnessed Anthropic's Claude AI to develop autonomous lethal drones and enhance their hacking and influence operations. This alarming revelation comes from a report released by the U.S.-based AI company, which underscores the growing concern that nations like Russia, China, and Iran are exploiting American technology to target global interests.
According to Anthropic, a Russian group known as GTG-27005 has utilized Claude to create an AI model that can operate on single-board computers, enabling drones to autonomously select targets and execute detonation commands without human intervention. Although this model has not yet been deployed in the field, the group has conducted real hardware-in-loop testing, indicating a serious intent to operationalize these capabilities. In response, Anthropic has banned the group and implemented additional safeguards based on its findings.
The report also highlights attempts by Chinese and Yemeni groups to use Claude for designing conventional weapons, efforts that Anthropic claims to have disrupted. Furthermore, the report details how foreign entities have employed Claude in sophisticated data theft and disinformation campaigns. These operations have included gathering intelligence on targets, exploiting online vulnerabilities, and creating malware to steal sensitive information.
Notably, the report indicates that these groups have constructed networks of fake social media profiles and entire news sites to disseminate misinformation. The tactics mirror those of Russian influence campaigns, which often involve the use of fabricated news outlets and journalist profiles to propagate false narratives about NATO activities in the Baltics. The integration of AI tools has significantly streamlined these operations, allowing for a greater volume of disinformation to be produced and making it more convincing to local audiences.
Another group, GTG-84005, has reportedly used AI to access census and public information, enabling them to tailor their messaging campaigns to specific demographics in Malaysia. This group has been known to launder Russian and Chinese state media as independent reporting, further complicating the information landscape.
Russia's Use of AI in Disinformation Efforts
The findings from Anthropic come shortly after OpenAI disclosed that Russian actors had utilized ChatGPT to fabricate a fake expert community aimed at promoting specific talking points to international audiences. While both Anthropic and OpenAI have managed to disrupt some of these schemes, the report warns that such operations are likely to persist as Russia continues to leverage AI to amplify content that aligns with its narratives.
In addition, the report notes that Russia is capitalizing on the curiosity of young people regarding AI, offering internships and training programs to foster a new generation of individuals who can disseminate Russian disinformation. This trend raises significant concerns about the potential for AI to empower authoritarian regimes in distorting global perceptions of geopolitical events.
Case Closed
Experts have pointed out that authoritarian regimes are quick to adopt AI technologies to manipulate reality in their favor. This shift is reminiscent of past U.S. government initiatives aimed at countering foreign disinformation, which have been undermined in recent years. The Trump administration's decision to dissolve the FBI's Foreign Influence Task Force and shutter the Counter Foreign Information Manipulation and Interference hub has left a void in the U.S. response to these threats.
The implications of these developments are profound, as the 2025 White House AI Action Plan has been criticized for omitting references to misinformation, particularly when it originates from foreign sources. This oversight could hinder efforts to combat the growing influence of AI-generated disinformation in the geopolitical arena.
