Researchers from Ben Gurion University have highlighted the alarming vulnerability of AI chatbots, revealing that they can be easily manipulated to produce harmful and illegal information. The study points to the emergence of “dark LLMs,” intentionally designed without safety measures, which can assist with cybercrime and fraud. Despite alerting major LLM providers about a universal jailbreak that compromises various chatbots, the researchers reported a lackluster response, underscoring the urgent need for enhanced safety protocols in AI systems.
