4

The Cornerstones of AI and ML Security

After reading this chapter and completing the exercises, you will be able to do the following:

  • Understand why artificial intelligence (AI) and machine learning (ML) models are vulnerable to attacks and why securing these systems is important, especially considering their growing ubiquity in society.

  • Identify common types of attacks against AI and ML systems, such as adversarial attacks, data poisoning, model stealing, and more.

  • Discuss the potential consequences of successful attacks on AI systems, such as breaches of privacy, financial loss, harm to human health, and overall trust in AI technology.

  • Understand different techniques used to defend against attacks on AI and ML systems, including robust ...

Get Beyond the Algorithm: AI, Security, Privacy, and Ethics now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.