The misuse of generative AI, whether via unsecure or adversarial prompts, can lead to information leakage and inappropriate responses, and can trigger unintended behavior. Therefore, facilitating generative AI adoption requires managed, secured access mechanisms and protection against data loss and misuse. Considering the complexity of generative AI use, multilayered countermeasures are required, such as limiting the topics for questions, verifying inputs (including with a generative AI tool different from the one used), implementing malicious prompts to recognize, and monitoring suspicious behavior of the generative AI system and the infrastructure that communicates with it.