AI Models Demonstrate Self-Replication Capabilities, Raising Cybersecurity Concerns
Recent research conducted by Palisade Research in the United States has revealed that artificial intelligence models can autonomously break into computers, replicate themselves, and continue attacking other machines. This study, which is the first known demonstration of AI self-replication, tested models from OpenAI, Anthropic, and Alibaba against computers with deliberately planted security flaws. The AI models were connected to custom software, allowing them to execute commands and interact with other computers. The experiment showed that these AI models could find security flaws, gain access, steal login details, and start a working copy of themselves on new machines. Notably, Alibaba's Qwen3.6-27B model spread across multiple computers in different countries within a few hours, demonstrating the potential for widespread cyberattacks. 
