SecureChat acts as an intermediary between users and large language models (LLMs) to detect and block sensitive data from being transmitted to generative-AI chatbots. When a user inputs a prompt containing sensitive information, SecureChat substitutes the sensitive data with synthetic but similar information before sending it to the LLM. Upon receiving the response from the LLM, which includes the synthetic data, SecureChat reconstructs the response with the original sensitive information for presentation to the user, ensuring sensitive data remains within organizational boundaries.
