Trust, privacy and transparency have always been at the core of Trustworthy, and our work with artificial intelligence (AI) is no different.
Trustworthy does not use data from customers to train AI models.
Trustworthy uses AI in a way that is "memoryless" - the model does not change or update when we interact with it, and your information is not learned by the model. While there are existing AI-powered services that do update or learn from your information, this is not something that is innate to using AI models. The processes of "reasoning" (i.e. asking the model questions about a given document) and "learning" (i.e. instructing the model to remember details of a given document) are entirely separate. We use AI models exclusively for their reasoning capabilities, and do not update based on any data provided by our users. It is not possible for "learning" to occur unintentionally.
In addition, when working with third-party models running on the Trustworthy private cloud, Trustworthy and its partners agree on robust protections to safeguard the data privacy and security of our customers and users.