Cisco Talos finds hackers using AI tools to build malware
Cisco's Talos team has spotted hackers using popular AI tools like Claude Code, Codex, Cursor, and Gemini to build malware and automate cyberattacks.
After digging through prompt histories and chat logs accidentally exposed online by the hackers, researchers found hackers are finding clever ways around the safety features built into these AIs, raising fresh worries about how easily AI can be misused for cybercrime.
Hackers use simple tricks, stolen tokens
Turns out, most of these tricks aren't high-tech at all. Hackers simply restart chats, participate in an authorized 'ethical hacking' competition, or claim admin privileges to fool the system.
Some even use stolen API tokens and compromised accounts to run their operations on corporate compute power rather than paying for their own infrastructure.
Cisco warns that as AI tools get more advanced, it's getting tougher for security teams to keep up.