Recently, the field of artificial intelligence (AI) security has once again attracted attention. According to Wired, researchers from Cornell University, the Israel Institute of Technology, and Intuit have jointly developed a new type of AI worm called "Morris II". This worm has powerful data stealing capabilities and the potential to breach security defenses, posing a serious threat to AI platforms such as ChatGPT and Gemini.

The Morris II worm is designed to spread through AI-driven email clients. Once released, it generates self-replicating prompts within the client, quickly occupying system resources and causing the email client to continuously forward emails, thus rapidly spreading across the network. This mechanism not only cripples email systems but also allows the worm to collect and transmit sensitive user data.

It is worth noting that Morris II is different from traditional worm viruses in that it utilizes AI models for replication and evolution, enabling it to adapt to different security environments and continuously enhance its attack capabilities. The research team demonstrated in experiments how this worm easily breaches the security defenses of AI platforms like ChatGPT and Gemini, and steals critical information stored within them.

Industry experts have expressed concerns about this. They believe that as AI technology becomes more widespread and applied, such worm viruses specifically targeting AI systems may become more complex and difficult to prevent. This also serves as a reminder to companies and developers to strengthen their focus and investment in AI security while pursuing powerful AI functionality.

Currently, the research team has publicly released detailed information about Morris II and called on the AI security community to collaborate on countermeasures to prevent potentially greater damage from such worm viruses in the future. With the rapid development of AI technology, ensuring the security and stability of AI systems has become an urgent problem to be solved.