Amazon security engineer hacks PC accessories with ease using Claude Opus (www.techradar.com)

A security engineer at Amazon, Chaz Schlarp, recently demonstrated the alarming potential of AI in hacking peripheral devices like webcams and microphones using Claude Opus 5. In a blog post detailing his experiments, Schlarp revealed how he modified firmware across various devices — including an Insta360 webcam and an Elgato video capture stick — in as little as 13 hours with minimal human oversight. This ease of use raises significant concerns for the AI/ML community regarding the vulnerability of widely used peripherals, as the very same techniques could be exploited by malicious actors to compromise systems stealthily. Scharp's findings indicate that most peripherals lack robust firmware integrity protections, facilitating unauthorized modifications. He showcased practical examples, such as removing annoying notifications from a monitor and disabling indicator lights on cameras and microphones, which could be leveraged for malicious surveillance. As AI reduces the barriers to executing complex firmware attacks, it creates a daunting future where IoT devices may become entry points for widespread cyber threats. Schlarp warns that this evolving capability could lead to the emergence of AI-powered malware capable of self-replication, making it imperative for the security community to reassess device integrity and user trust in the context of increasing AI sophistication.