Beijing's AI Army Hacks the Planet! Anthropic Sounds Alarm, US Fights Back in Cyber Showdown

This is your Cyber Sentinel: Beijing Watch podcast. Hey listeners, Ting here with your Beijing Watch. Grab your cyber helmets—this week’s Chinese activity pulse is off the charts! Let’s jump straight in: If you’re tracking new attack methodologies, you’ll want to know about the game-changing move by a Chinese state-sponsored hacking group: they took AI to the offensive. Anthropic’s security team announced that in September, attackers weaponized its Claude Code tool to run an almost fully automated cyberattack against multiple US tech firms and government agencies. Anthropic says these hackers only needed to steer the AI for about 10 to 20 percent of the work—the rest, all handled by generative code[1]. Imagine your pentester becoming the laziest human in the world because AI is doing the drudgework. So much for job security. Let’s talk industries: Transportation was hit hard—last Friday, DragonForce, a ransomware group with reported China ties, claimed responsibility for a breach at Barr Trucking Inc., threatening to leak sensitive data unless the company negotiates. That’s not a random pick; logistics and movement data are gold in a US-China contest, as control and disruption here can ripple through supply chains[4]. Energy, healthcare, and semiconductor manufacturing remain major targets. According to Ermer & Suter, government assessments confirm continued campaigns by Chinese actors aiming at communications, transportation, water, and power. Former FBI Director Christopher Wray has called out that these intrusions seek real operational harm, not just data theft—think: blackout, water contamination, or choking comms if a broader conflict broke out[2][3]. Now, attribution: It’s not just signals and whispers. Positive Technologies in Russia, of all places, published analysis confirming years-long advanced persistent threat operations by APT31, the notorious Chinese group, stealing from Russian and Western IT contractors[1]. Anthropic’s attribution of the AI-enabled operation directly to a Chinese state group raises global alarm—this is a leap in automation and plausible deniability. Also, those plucky Five Eyes countries (US, UK, Australia, New Zealand, Canada) are tightening the noose with new advisories, more sanctions, and—this is rare—public warnings that China’s Ministry of State Security and affiliated think tanks are stepping beyond industrial espionage into front-line operational attacks[7]. International response? It’s escalating fast. The US, UK, and Australia just sanctioned several Chinese infrastructure and tech companies seen as “dual-use” risk, and the Committee on Foreign Investment in the United States is blocking more takeovers of critical assets. The semiconductor world is a flashpoint, with Arizona’s chip factories, especially TSMC and Intel, now subject to stepped-up federal protection after evidence that Chinese hardware from Bitmain could be leveraged as a remote backdoor[1]. The SEC, FCC, and CISA are all shifting regu This content was created in partnership and with the help of Artificial Intelligence AI.