ChatGPT may share user inputs, security experts say


Personal data shared during interactions with AI chatbots may be vulnerable to leaks, putting both individual and company information at risk. According to Professor Justin Cappos, “This is one of the dark little secrets that a lot of companies don’t really understand as they use AI tools…that the training information that you put into your AI tool that creates this model that does your helpdesk process or whatever else, if you put anything sensitive in there, that information very much can be leaked back out.”