Today's AI can beat humans at Jeopardy, chess, recognizing faces and diagnosing medical conditions. As of last Fall it can write malware, too. In fact, it can write an entire attack chain: phishing emails, macros, reverse shells, you name it. What do we do now?
