Weaponizing GANs for Deepfakes and Adversarial Attacks
In the previous chapter, we built a foundational understanding of Generative AI and covered GANs in detail. Generative AI brings new challenges and GANs create a form of AI that can be used to accelerate adversarial attacks.
In this chapter, we will cover how GANs can be used in adversarial attacks. We will cover the following topics:
- Using GANs for deepfakes and deepfake detection
- Using GANs in cyberattacks such as evading face verification and malware detection, cracking passwords, compromising biometric authentication, cryptographic attacks, and generating payloads for traditional application attacks.
- Using GANs in pure adversarial AI attacks to create payloads to assist other forms of adversarial AI
- Defenses and mitigations to secure GANs, mitigate their use in adversarial attacks, and safeguard their use
Let’s start with the use of GANs for deepfakes.
