On "Google Gemini Can Now Control Your Computer. Hackers Are Already Targeting AI Agents"

Google just handed every SEO a new attack surface to worry about. If agentic AI is going to click through your site and execute actions, malicious actors will absolutely weaponize that—hidden redirects, fake forms, social engineering traps designed to trick AI into doing things it shouldn't. This isn't a theoretical threat; it's the obvious next move once these agents go mainstream. For anyone building sites that AI will actually navigate and interact with (not just crawl), you need to think about adversarial input the way you'd think about injection attacks, because the stakes just shifted from "corrupted search results" to "rogue AI following malicious instructions on your domain."