The interplay between the use of AI tools and the survival of legal privilege protection has not been explored in detail; importantly, there is no detailed guidance available from relevant professional bodies, authorities or the courts. Therefore, general principles apply. In our view, if AI tools are used as a supporting tool for legal analyses or document creation, the input and output should remain protected by legal privilege, provided that the input and output data are not disclosed to the AI provider or third parties and the processing is carried out under the lawyer's control. In contrast, disclosing input or output data in connection with the use of cloud-based AI services (e.g., ChatGPT or Copilot) without a contractual guarantee of confidentiality from the service provider could be seen as a disclosure to a large number of third parties, which could eventually lead to a loss of protection.