AI War Begins: China and Russia Utilising American-Made Weapons
China and Russia are reportedly utilising artificial intelligence (AI) from the United States to conduct various dangerous activities, ranging from cyberattacks and digital surveillance to weapons development.
These findings were disclosed by the US AI company Anthropic in a recent report cited by Politico. Anthropic discovered that criminal hacking groups, Chinese security apparatuses, Russian-linked spies, and even weapons manufacturers in Yemen have been using fake accounts, VPNs, and intermediary services to access the Claude AI assistant, despite restricted access.
Over the past eight months, these actors are said to have used Claude to automate cyberattacks, spread propaganda, conduct mass surveillance, and design sophisticated military hardware.
Anthropic revealed that Chinese and Iranian actors used Claude to analyse social media content to identify potential dissidents and politically sensitive information. Meanwhile, a Russia-based actor attempted to use Claude to develop drones capable of performing autonomous kamikaze attacks.
Anthropic also intercepted activities from a group in North Yemen using Claude for the development of guided rockets, multi-stage ballistic missiles, and hypersonic glide vehicles.
Furthermore, the company identified five instances where scientists in foreign nations used Claude to research dangerous pathogens, including studies on deadly viruses and highly pathogenic avian influenza. Anthropic did not disclose the specific countries or identities of the researchers, as it could not yet confirm whether the research was for legitimate purposes or biological weapons programmes.
“We are not trying to be hyperbolic here. We simply want to show the world how technology can actually be misused today,” said Jacob Klein, Anthropic’s Head of Threat Research.
Klein emphasised that these cases did not involve Anthropic’s most powerful AI model, Claude Mythos. However, he warned that several instances of misuse that were previously considered merely hypothetical have now become a reality.
Anthropic stated that it has halted all identified cases, reported them to relevant governments and industry groups, and strengthened Claude’s security systems.
This report emerges as Washington and Silicon Valley face increasing pressure to enhance AI security. One Anthropic researcher even resigned this week in protest against the industry race to develop increasingly powerful AI models.
A senior Anthropic scientist subsequently wrote that many AI safety staff sincerely believe there is approximately a 10% chance that AI could kill all humans within the next decade.
This statement has triggered a series of urgent calls from US Congressional members on Wednesday to implement more aggressive new safety regulations for the AI industry.
Source: jawawa.id (https://jawawa.id/newsitem/ai-war-begins-china-and-russia-utilising-american-made-weapons-1789109852)
Originally published by CNBC Indonesia (https://www.cnbcindonesia.com/tech/20260911122947-37-767175/perang-ai-dimulai-china-dan-rusia-pakai-senjata-buatan-amerika), translated from Indonesian