AI is changing how users interact with applications. An AI assistant may read information, trigger actions, call APIs, or perform tasks on behalf of a user. That creates a new security problem for Android developers.

We already think about protecting apps from malicious users, compromised devices, insecure networks, and reverse engineering. But what happens when an AI system can interact with our application and perform actions for the user?

In this session, we will look at the security challenges that come with AI powered and agent driven Android applications. We will explore how to protect sensitive data, verify identity, control what an AI system is allowed to do, and define clear boundaries between the user, the AI, the application, and the backend.

The focus will be on practical Android security rather than AI theory. We will look at real scenarios and discuss how existing Android security mechanisms can be combined with authorization, least privilege, validation, and human approval for sensitive actions.

Key Takeaways

• Understand how AI and agent based workflows introduce new security risks for Android applications

• Learn how to protect sensitive application data from unnecessary AI access and exposure

• Understand the difference between authentication and authorization when AI performs actions on behalf of a user

• Design app capabilities with least privilege so AI systems can only perform actions they are explicitly allowed to perform

• Learn how to validate and protect high risk actions instead of blindly trusting AI generated requests.