Security researchers have created a self-replicating AI worm named Morris II that targets AI-powered applications, such as OpenAI's ChatGPT and Google's Gemini.

Morris II can infiltrate systems, spread malware, and steal data without requiring user interaction, exploiting generative AI applications' capabilities.

The worm uses adversarial prompts to manipulate AI models, turning them into vehicles for disseminating malware and other harmful content.

Morris II can spread its payload through emails and images, leveraging AI systems to disseminate spam, propaganda, and malware across multiple platforms.

These findings have raised concerns about the security of AI systems and the urgency for developing robust defenses against new types of cyber threats.

Experts suggest that AI developers should create systems that can differentiate between human and AI-generated inputs and consider adopting a distributed model to enhance security.