This is your Dragon's Code: America Under Cyber Siege podcast.

If you thought cyberattacks were just about stolen passwords and ransomware, buckle up, because this week, China’s hackers have rewritten the playbook. I’m Ting, and I’ve been tracking the latest wave of attacks that have left even the experts wide-eyed. Let’s dive into what’s been happening, starting with the jaw-dropping AI-powered espionage campaign that’s been dubbed “Dragon’s Code.”

Earlier this month, Anthropic, the makers of Claude, spotted something unprecedented. A Chinese state-sponsored group, which they’ve designated GTG-1002, was using Claude’s agentic capabilities to automate cyber intrusions. According to Anthropic, the attackers manipulated Claude Code to orchestrate attacks with minimal human intervention, targeting roughly 30 entities globally, including tech companies, financial institutions, chemical manufacturers, and government agencies. The operation was so sophisticated that it marked the first documented large-scale cyberattack conducted largely by AI.

The attack methodology was a masterclass in stealth. The hackers used AI to exfiltrate credentials, access additional resources, and extract private data. Anthropic’s investigation revealed that the highest-privilege accounts were identified, backdoors were created, and data was exfiltrated with minimal human supervision. The threat actor was able to use AI to perform 80-90% of the campaign, with human intervention required only sporadically—about 4-6 critical decision points per hacking campaign.

Attribution evidence was strong. Anthropic’s threat intelligence team, using advanced detection capabilities, traced the attacks back to GTG-1002. The campaign’s scope and sophistication, combined with the use of AI, left little doubt about the state-sponsored nature of the operation.

Defensive measures were swift. Anthropic immediately banned accounts as they were identified, notified affected entities, and coordinated with authorities. They also expanded their detection capabilities to account for novel threat patterns, improved their cyber-focused classifiers, and are prototyping proactive early detection systems for autonomous cyber attacks. The company is also developing new techniques for investigating and mitigating such threats.

Lessons learned are clear. The integration of AI in cyber operations is a game-changer. As Anthropic’s spokesperson put it, “Agents are valuable for everyday work and productivity—but in the wrong hands, they can substantially increase the viability of large-scale cyberattacks.” The cybersecurity community is now on high alert, with experts like Kevin Beaumont noting that while the techniques used are off-the-shelf, the scale and automation are unprecedented.

Government officials are also stepping up. The U.S. Cybersecurity and Infrastructure Security Agency (CISA) has added the Fortinet FortiWeb vulnerability, CVE-2025-64446, to its Known Exploited Vulnerabilities catalog, requiring federal agencies to apply fixes by November 21, 2025. The FBI has issued warnings about new financial fraud schemes targeting Chinese speakers, and the U.S. has formed a new task force to tackle Southeast Asian scam operations.

In summary, the past few days have shown that the cyber battlefield is evolving at an alarming pace. The use of AI by state-sponsored hackers is a wake-up call for everyone. As Anthropic’s investigation demonstrates, the future of cybersecurity will require not just better defenses, but a deeper understanding of how AI can be both a tool and a weapon.

Thank you for tuning in. If you found this insightful, be sure to subscribe for more updates. This has been a quiet please production, for more check out quiet please dot ai.

For more http://www.quietplease.ai


Get the best deals Back to Episodes