PDC Bank is working on creating an AI application that enables customers to send SMS to the AI application to allow banking activities from their registered ID. Jane, the project engineer, has taken bank customer data from the last few years from the server and is using it to train the ML to recognize and authenticate actual users and to ensure unauthorized users are barred from entering the application. Suppose the AI application has been compromised, and the reason has been identified as compromised data being used to improve the ML accuracy. What kind of attack is the PDC Bank application subjected to?