Leading cybersecurity expert Dan Lahav, who tests the security of AI models, reveals exclusively to Sky News a major problem with AI and password generation. The issue affects some of the most used AI models such as ChatGPT, Claude and Gemini. Sky News Technology Correspondent Rowland Manthorpe explains why AI-generated passwords hide a major security flaw and also discusses with Dan the broader issue of AI and cybersecurity, particularly in the context of AI systems being able to find and