Part 4: Common Problems When Applying AI in Cybersecurity

In this part, we look back at the AI approaches and describe common problems that need to be solved in order to make them successful. Based on our experience, we recollect common pitfalls, structure the problems that need special attention, and describe approaches to overcome those problems. We start with data quality as a prerequisite for successful AI application, and then move forward to the need for a proper understanding of the limitations of statistical approaches, as well as proper evaluation and monitoring. We finish by describing the challenges of a changing environment, and the rising need for responsible AI application.

This part has the following chapters:

• Chapter 14, Data Quality and its Usage in the AI and LLM Era
• Chapter 15, Correlation, Causation, Bias, and Variance
• Chapter 16, Evaluation, Monitoring, and Feedback Loop
• Chapter 17, Learning in a Changing and Adversarial Environment
• Chapter...