Matthew RosenquistFounder & CISO Advisor, Cybersecurity Insights · CISO, Cybersecurity Strategist, & Industry Advisor

Researchers Give Birth to the First GenAI Worm

4 Mar 2024 · 1 min

Watch on YouTube

1:16

In this brief cybersecurity commentary, the speaker explains that researchers have created the first generative AI worm, named Morris2, which can steal data, self‑propagate, and exploit vulnerabilities in AI‑assisted email platforms using adversarial AI techniques. Although the worm is a research prototype, it successfully demonstrates functional malicious capabilities, highlighting a new class of emerging GenAI‑driven threats. The speaker urges the cybersecurity community to accelerate proactive risk management for generative AI, balancing its benefits with security, and invites viewers to follow for more insights.

Chapters

Researchers just created the first generative AI worm that can conduct malicious activities, steal data, and most interestingly, automatically self-propagate itself to others. The worm, dubbed Morris2, exploits vulnerabilities in AI-assisted email platforms and uses adversarial AI techniques to sneak in instructions to the victim's email system. And although this is just research, they actually got the worm to work, and it is a glimpse of the emerging threats that we will all face. Now, the cybersecurity community knew this type of capability was coming, but now we need to move faster to proactively manage the cyber risks of GenAI because it is evolving so fast and being adopted very quickly. If we do it right, we can all benefit from the great capabilities of AI while being reasonably secure, but it will take effort and forethought. Follow me for more cybersecurity insights.