GoKawiilThreat actors in China, Iran, and Russia have been moving away from simple prompt-and-reply chatbots to more autonomous systems, a new report from Google has found.
Silicon Valley’s AI Agent Push Has Been Paying Off—for Cybercriminals
Why This Matters
Google's new threat-intelligence report finds state-linked actors in China, Iran and Russia shifting from using chatbots for one-off prompts to deploying more autonomous AI agents in their operations. That mirrors the same agentic push Silicon Valley is selling to enterprises, meaning the industry's flagship capability is being adopted just as quickly by adversaries. For vendors and customers, it raises hard questions about whether safeguards built for chatbots are adequate for systems that act on their own.
Key Takeaways
- Google reports state-linked threat actors in China, Iran and Russia are adopting autonomous AI systems, not just prompt-and-reply chatbots.
- Agentic AI, the industry's biggest current bet, is proving equally useful to attackers.
- Guardrails designed for conversational models may not cover AI that takes multi-step actions independently.
Worth a Look
YubiKey 5C NFC Security Key — When AI-powered attackers are automating phishing and account takeovers, a hardware security key is one of the strongest defenses you can add to your logins. The YubiKey 5C NFC works with major services over USB-C or by tapping your phone, so your accounts need physical possession of the key, not just a stealable password or code.
See YubiKey 5C NFC Security Key on Amazon → Affiliate link — we may earn a commission on purchases, at no extra cost to you. Product picked by AI based on this article; it is not a tested recommendation.Get alerts for these topics