Most of recent issues, including this incident, happened not due to smart superintelligent "agents" taking over the world - chatbots and other text generators are about as intelligent amd powerful as a dead starfish - but due to the combined stupidity of the said chatbots amd lazy idiots who use them to hide their own incompetence and thus produce such embarassing mistakes. A few years ago, they would be fired for exposing secrets in plain text, but since their manager wanted an AI-Workflow...

How many can unequivocally be attributed to malicious AI?